Changes between Version 80 and Version 81 of WKTRaster/SpecificationWorking03


Ignore:
Timestamp:
May 16, 2011, 11:54:53 AM (13 years ago)
Author:
Bborie Park
Comment:

Move statistics functions to FV.16

Legend:

Unmodified
Added
Removed
Modified

  • WKTRaster/SpecificationWorking03

    v80 v81  
    100100
    101101''' Open Question: ''' Should the function fail if an index is invalid?  How should this work when providing more than one indices to the function?
    102 
    103 ----
    104 
    105 '''ST_SummaryStats(raster, nband) -> record'''[[BR]]
    106 This is the core function that gets the summary statistics (# of values, mean, standard deviation, minimum value, maximum value) of a specified raster's band.  It is this function that ST_Mean, ST_StdDev and ST_MinMax calls for their appropriate values.
    107 
    108 1. ST_SummaryStats(rast raster, nband int, hasnodata boolean) -> record
    109 
    110   returns one record of five columns (count, mean, stddev, min, max)
    111 
    112   nband: index of band
    113 
    114   hasnodata: if FALSE, any pixel who's value is nodata is ignored
    115 
    116 {{{
    117 ST_SummaryStats(rast, 1, FALSE)
    118 }}}
    119 
    120 2. ST_SummaryStats(rast raster, nband int) -> record
    121 
    122   assumes hasnodata = FALSE
    123 
    124 {{{
    125 ST_SummaryStats(rast, 2)
    126 }}}
    127 
    128 3. ST_SummaryStats(rast raster, hasnodata boolean) -> record
    129 
    130   assumes nband = 1
    131 
    132 {{{
    133 ST_SummaryStats(rast, TRUE)
    134 }}}
    135 
    136 4. ST_SummaryStats(rast raster) -> record
    137 
    138   assumes nband = 1 and hasnodata = FALSE
    139 
    140 {{{
    141 ST_SummaryStats(rast)
    142 }}}
    143 
    144 Due to the time it may take to do on-the-fly calculation of summary stats for large rasters (say 10000 x 10000), an alternative that sacrifices accuracy for speed is required.  The following functions sample a percentage of the raster in a methodical randomized manner.  The algorithm used for sampling is...
    145 
    146 1. select the larger dimension of the width and height.  compute the number of pixels to sample in each "row" of the larger dimension
    147 
    148 2. pick pixels from each "row" of the larger dimension in an incremental rolling manner where each increment is randomly determined.
    149 
    150 The set of ST_ApproxSummaryStats functions are:
    151 
    152 1. ST_ApproxSummaryStats(rast raster, nband int, hasnodata boolean, sample_percent double precision) -> record
    153 
    154   sample_percent: a value between 0 and 1 indicating the percentage of the raster band's pixels to consider
    155 
    156 {{{
    157 ST_ApproxSummaryStats(rast, 3, FALSE, 0.1)
    158 
    159 ST_ApproxSummaryStats(rast, 1, TRUE, 0.5)
    160 }}}
    161 
    162 2. ST_ApproxSummaryStats(rast raster, nband int, sample_percent double precision) -> record
    163 
    164   assumes that nband = 1
    165 
    166 {{{
    167 ST_ApproxSummaryStats(rast, 2 0.01)
    168 
    169 ST_ApproxSummaryStats(rast, 4, 0.025)
    170 }}}
    171 
    172 3. ST_ApproxSummaryStats(rast raster, hasnodata boolean, sample_percent double precision) -> record
    173 
    174   assumes that nband = 1
    175 
    176 {{{
    177 ST_ApproxSummaryStats(rast, FALSE, 0.01)
    178 
    179 ST_ApproxSummaryStats(rast, TRUE, 0.025)
    180 }}}
    181 
    182 4. ST_ApproxSummaryStats(rast raster, sample_percent double precision) -> record
    183 
    184   assumes that nband = 1 and hasnodata = FALSE
    185 
    186 {{{
    187 ST_ApproxSummaryStats(rast, 0.25)
    188 }}}
    189 
    190 5. ST_ApproxSummaryStats(rast raster) -> record
    191 
    192   assumes that nband = 1, hasnodata = FALSE and sample_percent = 0.1
    193 
    194 {{{
    195 ST_ApproxSummaryStats(rast)
    196 }}}
    197 
    198 The situation arises where the summary statistics of a coverage table is required.  As the coverage may be large (tens of gigabytes of memory or larger), the following functions are provided to permit an incremental computation of the summary statistics.
    199 
    200 1. ST_SummaryStats(rastertable text, rastercolumn text, nband int, hasnodata boolean) -> record
    201 
    202   rastertable: name of table with raster column
    203 
    204   rastercolumn: name of column of data type raster
    205 
    206 {{{
    207 ST_SummaryStats('tmax_2010', 'rast', 1, FALSE)
    208 
    209 ST_SummaryStats('precip_2011', 'rast', 1, TRUE)
    210 }}}
    211 
    212 2. ST_SummaryStats(rastertable text, rastercolumn text, nband int) -> record
    213 
    214     hasnodata = FALSE
    215 
    216 {{{
    217 ST_SummaryStats('tmax_2010', 'rast', 1)
    218 }}}
    219 
    220 3. ST_SummaryStats(rastertable text, rastercolumn text, hasnodata boolean) -> record
    221 
    222     nband = 1
    223 
    224 {{{
    225 ST_SummaryStats('precip_2011', 'rast', TRUE)
    226 }}}
    227 
    228 4. ST_SummaryStats(rastertable text, rastercolumn text) -> record
    229 
    230     nband = 1 and hasnodata = FALSE
    231 
    232 {{{
    233 ST_SummaryStats('tmin_2009', 'rast')
    234 }}}
    235 
    236 Variations for ST_ApproxSummaryStats are:
    237 
    238 1. ST_ApproxSummaryStats(rastertable text, rastercolumn text, nband int, hasnodata boolean, sample_percent double precision) -> record
    239 
    240 {{{
    241 ST_ApproxSummaryStats('tmax_2010', 'rast', 1, FALSE, 0.5)
    242 
    243 ST_ApproxSummaryStats('precip_2011', 'rast', 1, TRUE, 0.2)
    244 }}}
    245 
    246 2. ST_ApproxSummaryStats(rastertable text, rastercolumn text, nband int, sample_percent double precision) -> record
    247 
    248     hasnodata = FALSE
    249 
    250 {{{
    251 ST_ApproxSummaryStats('tmax_2010', 'rast', 1, 0.5)
    252 
    253 ST_ApproxSummaryStats('precip_2011', 'rast', 1, 0.2)
    254 }}}
    255 
    256 3. ST_ApproxSummaryStats(rastertable text, rastercolumn text, hasnodata boolean, sample_percent double precision) -> record
    257 
    258     nband = 1
    259 
    260 {{{
    261 ST_ApproxSummaryStats('tmax_2010', 'rast', FALSE, 0.5)
    262 
    263 ST_ApproxSummaryStats('precip_2011', 'rast', TRUE, 0.2)
    264 }}}
    265 
    266 4. ST_ApproxSummaryStats(rastertable text, rastercolumn text, sample_percent double precision) -> record
    267 
    268     nband = 1 and hasnodata = FALSE
    269 
    270 {{{
    271 ST_ApproxSummaryStats('tmax_2010', 'rast', 0.5)
    272 
    273 ST_ApproxSummaryStats('precip_2011', 'rast', 0.2)
    274 }}}
    275 
    276 5. ST_ApproxSummaryStats(rastertable text, rastercolumn text) -> record
    277 
    278     nband = 1, hasnodata = FALSE and sample_percent = 0.1
    279 
    280 {{{
    281 ST_ApproxSummaryStats('tmax_2010', 'rast')
    282 
    283 ST_ApproxSummaryStats('precip_2011', 'rast')
    284 }}}
    285 
    286 The mean returned in the coverage functions (has rastertable and rastercolumn arguments) is a weighted mean of the means of each raster tile. The standard deviation returned is the cumulative standard deviation of all raster tiles.
    287 
    288 ----
    289 
    290 '''ST_Mean(raster, nband) -> double precision'''[[BR]]
    291 This function calls ST_SummaryStats and only returns the mean from that function.
    292 
    293 1. ST_Mean(rast raster, nband int, hasnodata boolean) -> double precision
    294 
    295   returns the mean as a double precision
    296 
    297   nband: index of band
    298 
    299   hasnodata: if FALSE, any pixel who's value is nodata is ignored
    300 
    301 {{{
    302 ST_Mean(rast, 1, FALSE)
    303 }}}
    304 
    305 2. ST_Mean(rast raster, nband int) -> double precision
    306 
    307   assumes hasnodata = FALSE
    308 
    309 {{{
    310 ST_Mean(rast, 2)
    311 }}}
    312 
    313 3. ST_Mean(rast raster, hasnodata boolean) -> double precision
    314 
    315   assumes nband = 1
    316 
    317 {{{
    318 ST_Mean(rast, TRUE)
    319 }}}
    320 
    321 4. ST_Mean(rast raster) -> double precision
    322 
    323   assumes nband = 1 and hasnodata = FALSE
    324 
    325 {{{
    326 ST_Mean(rast)
    327 }}}
    328 
    329 The set of ST_ApproxMean functions are:
    330 
    331 1. ST_ApproxMean(rast raster, nband int, hasnodata boolean, sample_percent double precision) -> double precision
    332 
    333   sample_percent: a value between 0 and 1 indicating the percentage of the raster band's pixels to consider
    334 
    335 {{{
    336 ST_ApproxMean(rast, 3, FALSE, 0.1)
    337 
    338 ST_ApproxMean(rast, 1, TRUE, 0.5)
    339 }}}
    340 
    341 2. ST_ApproxMean(rast raster, nband int, sample_percent double precision) -> double precision
    342 
    343   assumes that nband = 1
    344 
    345 {{{
    346 ST_ApproxMean(rast, 2 0.01)
    347 
    348 ST_ApproxMean(rast, 4, 0.025)
    349 }}}
    350 
    351 3. ST_ApproxMean(rast raster, hasnodata boolean, sample_percent double precision) -> double precision
    352 
    353   assumes that nband = 1
    354 
    355 {{{
    356 ST_ApproxMean(rast, FALSE, 0.01)
    357 
    358 ST_ApproxMean(rast, TRUE, 0.025)
    359 }}}
    360 
    361 4. ST_ApproxMean(rast raster, sample_percent double precision) -> double precision
    362 
    363   assumes that nband = 1 and hasnodata = FALSE
    364 
    365 {{{
    366 ST_ApproxMean(rast, 0.25)
    367 }}}
    368 
    369 5. ST_ApproxMean(rast raster) -> double precision
    370 
    371   assumes that nband = 1, hasnodata = FALSE and sample_percent = 0.1
    372 
    373 {{{
    374 ST_ApproxMean(rast)
    375 }}}
    376 
    377 The following functions are provided for coverage tables.
    378 
    379 1. ST_Mean(rastertable text, rastercolumn text, nband int, hasnodata boolean) -> double precision
    380 
    381   rastertable: name of table with raster column
    382 
    383   rastercolumn: name of column of data type raster
    384 
    385 {{{
    386 ST_Mean('tmax_2010', 'rast', 1, FALSE)
    387 
    388 ST_Mean('precip_2011', 'rast', 1, TRUE)
    389 }}}
    390 
    391 2. ST_Mean(rastertable text, rastercolumn text, nband int) -> double precision
    392 
    393     hasnodata = FALSE
    394 
    395 {{{
    396 ST_Mean('tmax_2010', 'rast', 1)
    397 }}}
    398 
    399 3. ST_Mean(rastertable text, rastercolumn text, hasnodata boolean) -> double precision
    400 
    401     nband = 1
    402 
    403 {{{
    404 ST_Mean('precip_2011', 'rast', TRUE)
    405 }}}
    406 
    407 4. ST_Mean(rastertable text, rastercolumn text) -> double precision
    408 
    409     nband = 1 and hasnodata = FALSE
    410 
    411 {{{
    412 ST_Mean('tmin_2009', 'rast')
    413 }}}
    414 
    415 Variations for ST_ApproxMean are:
    416 
    417 1. ST_ApproxMean(rastertable text, rastercolumn text, nband int, hasnodata boolean, sample_percent double precision) -> double precision
    418 
    419 {{{
    420 ST_ApproxMean('tmax_2010', 'rast', 1, FALSE, 0.5)
    421 
    422 ST_ApproxMean('precip_2011', 'rast', 1, TRUE, 0.2)
    423 }}}
    424 
    425 2. ST_ApproxMean(rastertable text, rastercolumn text, nband int, sample_percent double precision) -> double precision
    426 
    427     hasnodata = FALSE
    428 
    429 {{{
    430 ST_ApproxMean('tmax_2010', 'rast', 1, 0.5)
    431 
    432 ST_ApproxMean('precip_2011', 'rast', 1, 0.2)
    433 }}}
    434 
    435 3. ST_ApproxMean(rastertable text, rastercolumn text, hasnodata boolean, sample_percent double precision) -> double precision
    436 
    437     nband = 1
    438 
    439 {{{
    440 ST_ApproxMean('tmax_2010', 'rast', FALSE, 0.5)
    441 
    442 ST_ApproxMean('precip_2011', 'rast', TRUE, 0.2)
    443 }}}
    444 
    445 4. ST_ApproxMean(rastertable text, rastercolumn text, sample_percent double precision) -> double precision
    446 
    447     nband = 1 and hasnodata = FALSE
    448 
    449 {{{
    450 ST_ApproxMean('tmax_2010', 'rast', 0.5)
    451 
    452 ST_ApproxMean('precip_2011', 'rast', 0.2)
    453 }}}
    454 
    455 5. ST_ApproxMean(rastertable text, rastercolumn text) -> double precision
    456 
    457     nband = 1, hasnodata = FALSE and sample_percent = 0.1
    458 
    459 {{{
    460 ST_ApproxMean('tmax_2010', 'rast')
    461 
    462 ST_ApproxMean('precip_2011', 'rast')
    463 }}}
    464 
    465 The mean returned in the coverage functions (has rastertable and rastercolumn arguments) is a weighted mean of the means of each raster tile.
    466 
    467 ----
    468 
    469 '''ST_StdDev(raster, nband) -> double precision'''[[BR]]
    470 This function calls ST_SummaryStats and only returns the standard deviation from that function.
    471 
    472 1. ST_StdDev(rast raster, nband int, hasnodata boolean) -> double precision
    473 
    474   returns the standard deviation as a double precision
    475 
    476   nband: index of band
    477 
    478   hasnodata: if FALSE, any pixel who's value is nodata is ignored
    479 
    480 {{{
    481 ST_StdDev(rast, 1, FALSE)
    482 }}}
    483 
    484 2. ST_StdDev(rast raster, nband int) -> double precision
    485 
    486   assumes hasnodata = FALSE
    487 
    488 {{{
    489 ST_StdDev(rast, 2)
    490 }}}
    491 
    492 3. ST_StdDev(rast raster, hasnodata boolean) -> double precision
    493 
    494   assumes nband = 1
    495 
    496 {{{
    497 ST_StdDev(rast, TRUE)
    498 }}}
    499 
    500 4. ST_StdDev(rast raster) -> double precision
    501 
    502   assumes nband = 1 and hasnodata = FALSE
    503 
    504 {{{
    505 ST_StdDev(rast)
    506 }}}
    507 
    508 The set of ST_ApproxStdDev functions are:
    509 
    510 1. ST_ApproxStdDev(rast raster, nband int, hasnodata boolean, sample_percent double precision) -> double precision
    511 
    512   sample_percent: a value between 0 and 1 indicating the percentage of the raster band's pixels to consider
    513 
    514 {{{
    515 ST_ApproxStdDev(rast, 3, FALSE, 0.1)
    516 
    517 ST_ApproxStdDev(rast, 1, TRUE, 0.5)
    518 }}}
    519 
    520 2. ST_ApproxStdDev(rast raster, nband int, sample_percent double precision) -> double precision
    521 
    522   assumes that nband = 1
    523 
    524 {{{
    525 ST_ApproxStdDev(rast, 2 0.01)
    526 
    527 ST_ApproxStdDev(rast, 4, 0.025)
    528 }}}
    529 
    530 3. ST_ApproxStdDev(rast raster, hasnodata boolean, sample_percent double precision) -> double precision
    531 
    532   assumes that nband = 1
    533 
    534 {{{
    535 ST_ApproxStdDev(rast, FALSE, 0.01)
    536 
    537 ST_ApproxStdDev(rast, TRUE, 0.025)
    538 }}}
    539 
    540 4. ST_ApproxStdDev(rast raster, sample_percent double precision) -> double precision
    541 
    542   assumes that nband = 1 and hasnodata = FALSE
    543 
    544 {{{
    545 ST_ApproxStdDev(rast, 0.25)
    546 }}}
    547 
    548 5. ST_ApproxStdDev(rast raster) -> double precision
    549 
    550   assumes that nband = 1, hasnodata = FALSE and sample_percent = 0.1
    551 
    552 {{{
    553 ST_ApproxStdDev(rast)
    554 }}}
    555 
    556 The following functions are provided for coverage tables.
    557 
    558 1. ST_StdDev(rastertable text, rastercolumn text, nband int, hasnodata boolean) -> double precision
    559 
    560   rastertable: name of table with raster column
    561 
    562   rastercolumn: name of column of data type raster
    563 
    564 {{{
    565 ST_StdDev('tmax_2010', 'rast', 1, FALSE)
    566 
    567 ST_StdDev('precip_2011', 'rast', 1, TRUE)
    568 }}}
    569 
    570 2. ST_StdDev(rastertable text, rastercolumn text, nband int) -> double precision
    571 
    572     hasnodata = FALSE
    573 
    574 {{{
    575 ST_StdDev('tmax_2010', 'rast', 1)
    576 }}}
    577 
    578 3. ST_StdDev(rastertable text, rastercolumn text, hasnodata boolean) -> double precision
    579 
    580     nband = 1
    581 
    582 {{{
    583 ST_StdDev('precip_2011', 'rast', TRUE)
    584 }}}
    585 
    586 4. ST_StdDev(rastertable text, rastercolumn text) -> double precision
    587 
    588     nband = 1 and hasnodata = FALSE
    589 
    590 {{{
    591 ST_StdDev('tmin_2009', 'rast')
    592 }}}
    593 
    594 Variations for ST_ApproxStdDev are:
    595 
    596 1. ST_ApproxStdDev(rastertable text, rastercolumn text, nband int, hasnodata boolean, sample_percent double precision) -> double precision
    597 
    598 {{{
    599 ST_ApproxStdDev('tmax_2010', 'rast', 1, FALSE, 0.5)
    600 
    601 ST_ApproxStdDev('precip_2011', 'rast', 1, TRUE, 0.2)
    602 }}}
    603 
    604 2. ST_ApproxStdDev(rastertable text, rastercolumn text, nband int, sample_percent double precision) -> double precision
    605 
    606     hasnodata = FALSE
    607 
    608 {{{
    609 ST_ApproxStdDev('tmax_2010', 'rast', 1, 0.5)
    610 
    611 ST_ApproxStdDev('precip_2011', 'rast', 1, 0.2)
    612 }}}
    613 
    614 3. ST_ApproxStdDev(rastertable text, rastercolumn text, hasnodata boolean, sample_percent double precision) -> double precision
    615 
    616     nband = 1
    617 
    618 {{{
    619 ST_ApproxStdDev('tmax_2010', 'rast', FALSE, 0.5)
    620 
    621 ST_ApproxStdDev('precip_2011', 'rast', TRUE, 0.2)
    622 }}}
    623 
    624 4. ST_ApproxStdDev(rastertable text, rastercolumn text, sample_percent double precision) -> double precision
    625 
    626     nband = 1 and hasnodata = FALSE
    627 
    628 {{{
    629 ST_ApproxStdDev('tmax_2010', 'rast', 0.5)
    630 
    631 ST_ApproxStdDev('precip_2011', 'rast', 0.2)
    632 }}}
    633 
    634 5. ST_ApproxStdDev(rastertable text, rastercolumn text) -> double precision
    635 
    636     nband = 1, hasnodata = FALSE and sample_percent = 0.1
    637 
    638 {{{
    639 ST_ApproxStdDev('tmax_2010', 'rast')
    640 
    641 ST_ApproxStdDev('precip_2011', 'rast')
    642 }}}
    643 
    644 The standard deviation returned in the coverage functions (has rastertable and rastercolumn arguments) is the cumulative standard deviation of all raster tiles.
    645 
    646 ----
    647 
    648 '''ST_MinMax(raster, nband) -> record'''[[BR]]
    649 This function calls ST_SummaryStats and only returns the min and max values from that function.
    650 
    651 1. ST_MinMax(rast raster, nband int, hasnodata boolean) -> record
    652 
    653   returns the record (min double precision, max double precision)
    654 
    655   nband: index of band
    656 
    657   hasnodata: if FALSE, any pixel who's value is nodata is ignored
    658 
    659 {{{
    660 ST_MinMax(rast, 1, FALSE)
    661 }}}
    662 
    663 2. ST_MinMax(rast raster, nband int) -> record
    664 
    665   assumes hasnodata = FALSE
    666 
    667 {{{
    668 ST_MinMax(rast, 2)
    669 }}}
    670 
    671 3. ST_MinMax(rast raster, hasnodata boolean) -> record
    672 
    673   assumes nband = 1
    674 
    675 {{{
    676 ST_MinMax(rast, TRUE)
    677 }}}
    678 
    679 4. ST_MinMax(rast raster) -> record
    680 
    681   assumes nband = 1 and hasnodata = FALSE
    682 
    683 {{{
    684 ST_MinMax(rast)
    685 }}}
    686 
    687 The set of ST_ApproxMinMax functions are:
    688 
    689 1. ST_ApproxMinMax(rast raster, nband int, hasnodata boolean, sample_percent record) -> record
    690 
    691   sample_percent: a value between 0 and 1 indicating the percentage of the raster band's pixels to consider
    692 
    693 {{{
    694 ST_ApproxMinMax(rast, 3, FALSE, 0.1)
    695 
    696 ST_ApproxMinMax(rast, 1, TRUE, 0.5)
    697 }}}
    698 
    699 2. ST_ApproxMinMax(rast raster, nband int, sample_percent double precision) -> record
    700 
    701   assumes that nband = 1
    702 
    703 {{{
    704 ST_ApproxMinMax(rast, 2 0.01)
    705 
    706 ST_ApproxMinMax(rast, 4, 0.025)
    707 }}}
    708 
    709 3. ST_ApproxMinMax(rast raster, hasnodata boolean, sample_percent double precision) -> record
    710 
    711   assumes that nband = 1
    712 
    713 {{{
    714 ST_ApproxMinMax(rast, FALSE, 0.01)
    715 
    716 ST_ApproxMinMax(rast, TRUE, 0.025)
    717 }}}
    718 
    719 4. ST_ApproxMinMax(rast raster, sample_percent double precision) -> record
    720 
    721   assumes that nband = 1 and hasnodata = FALSE
    722 
    723 {{{
    724 ST_ApproxMinMax(rast, 0.25)
    725 }}}
    726 
    727 5. ST_ApproxMinMax(rast raster) -> record
    728 
    729   assumes that nband = 1, hasnodata = FALSE and sample_percent = 0.1
    730 
    731 {{{
    732 ST_ApproxMinMax(rast)
    733 }}}
    734 
    735 The following functions are provided for coverage tables.
    736 
    737 1. ST_MinMax(rastertable text, rastercolumn text, nband int, hasnodata boolean) -> record
    738 
    739   rastertable: name of table with raster column
    740 
    741   rastercolumn: name of column of data type raster
    742 
    743 {{{
    744 ST_MinMax('tmax_2010', 'rast', 1, FALSE)
    745 
    746 ST_MinMax('precip_2011', 'rast', 1, TRUE)
    747 }}}
    748 
    749 2. ST_MinMax(rastertable text, rastercolumn text, nband int) -> record
    750 
    751     hasnodata = FALSE
    752 
    753 {{{
    754 ST_MinMax('tmax_2010', 'rast', 1)
    755 }}}
    756 
    757 3. ST_MinMax(rastertable text, rastercolumn text, hasnodata boolean) -> record
    758 
    759     nband = 1
    760 
    761 {{{
    762 ST_MinMax('precip_2011', 'rast', TRUE)
    763 }}}
    764 
    765 4. ST_MinMax(rastertable text, rastercolumn text) -> record
    766 
    767     nband = 1 and hasnodata = FALSE
    768 
    769 {{{
    770 ST_MinMax('tmin_2009', 'rast')
    771 }}}
    772 
    773 Variations for ST_ApproxMinMax are:
    774 
    775 1. ST_ApproxMinMax(rastertable text, rastercolumn text, nband int, hasnodata boolean, sample_percent double precision) -> record
    776 
    777 {{{
    778 ST_ApproxMinMax('tmax_2010', 'rast', 1, FALSE, 0.5)
    779 
    780 ST_ApproxMinMax('precip_2011', 'rast', 1, TRUE, 0.2)
    781 }}}
    782 
    783 2. ST_ApproxMinMax(rastertable text, rastercolumn text, nband int, sample_percent double precision) -> record
    784 
    785     hasnodata = FALSE
    786 
    787 {{{
    788 ST_ApproxMinMax('tmax_2010', 'rast', 1, 0.5)
    789 
    790 ST_ApproxMinMax('precip_2011', 'rast', 1, 0.2)
    791 }}}
    792 
    793 3. ST_ApproxMinMax(rastertable text, rastercolumn text, hasnodata boolean, sample_percent double precision) -> record
    794 
    795     nband = 1
    796 
    797 {{{
    798 ST_ApproxMinMax('tmax_2010', 'rast', FALSE, 0.5)
    799 
    800 ST_ApproxMinMax('precip_2011', 'rast', TRUE, 0.2)
    801 }}}
    802 
    803 4. ST_ApproxMinMax(rastertable text, rastercolumn text, sample_percent double precision) -> record
    804 
    805     nband = 1 and hasnodata = FALSE
    806 
    807 {{{
    808 ST_ApproxMinMax('tmax_2010', 'rast', 0.5)
    809 
    810 ST_ApproxMinMax('precip_2011', 'rast', 0.2)
    811 }}}
    812 
    813 5. ST_ApproxMinMax(rastertable text, rastercolumn text) -> record
    814 
    815     nband = 1, hasnodata = FALSE and sample_percent = 0.1
    816 
    817 {{{
    818 ST_ApproxMinMax('tmax_2010', 'rast')
    819 
    820 ST_ApproxMinMax('precip_2011', 'rast')
    821 }}}
    822 
    823 ----
    824 
    825 '''ST_Histogram(raster, nband) -> set of records'''[[BR]]
    826 ST_Histogram and ST_ApproxHistogram provide methods to determine a raster's data distribution.
    827 
    828 The return of ST_Histogram and ST_ApproxHistogram is a set of records where each record is (min, max, count, proportion).
    829 
    830 ST_Histogram has the following variations.
    831 
    832 1. ST_Histogram(rast raster, nband int, hasnodata boolean, bins int, width double precision[], right boolean) -> set of records
    833 
    834   returns set of records of four columns (min, max, count, proportion)
    835 
    836   nband: index of band to process on
    837 
    838   hasnodata: if FALSE, any pixel who's value is nodata is ignored.
    839 
    840   bins: the number of categories/bins to have in the histogram. If NULL or value less than one, the number of categories will be auto-computed using Sturges' formula if the number of values >= 30 or Square-root choice if number of values < 30.
    841 
    842     http://en.wikipedia.org/wiki/Histogram#Mathematical_definition
    843 
    844   width: an array indicating the width of each category/bin. If the number of bins is greater than the number of widths, the widths are repeated. Example: 9 bins, widths are [a, b, c] will have the output be [a, b, c, a, b, c, a, b, c].
    845 
    846   right: compute the histogram from the right rather than from the left (default). This changes the criteria for evaluating a value x from [a, b) to (a, b].
    847 
    848 {{{
    849 ST_Histogram(rast, 2, FALSE, NULL, NULL, FALSE)
    850 
    851 ST_Histogram(rast, 1, TRUE, 100, NULL, FALSE)
    852 
    853 ST_Histogram(rast, 2, FALSE, NULL, ARRAY[100, 50], FALSE)
    854 
    855 ST_Histogram(rast, 2, FALSE, 9, ARRAY[1000], TRUE)
    856 
    857 ST_Histogram(rast, 2, FALSE, 20, ARRAY[100, 200, 300], TRUE)
    858 }}}
    859 
    860 2. ST_Histogram(rast raster, nband int, hasnodata boolean, bins int, right boolean) -> set of records
    861 
    862   parameter "width" is not specified thus resulting in all bins having the same widths
    863 
    864 {{{
    865 ST_Histogram(rast, 2, FALSE, 5, FALSE)
    866 }}}
    867 
    868 3. ST_Histogram(rast raster, nband int, hasnodata boolean, bins int) -> set of records
    869 
    870   the parameter "right" is removed and assumed to be FALSE
    871 
    872 {{{
    873 ST_Histogram(rast, 2, FALSE, 5)
    874 }}}
    875 
    876 4. ST_Histogram(rast raster, nband int, hasnodata boolean) -> set of records
    877 
    878   the parameter "bins" is removed and set to NULL.  The function will compute the number of bins to use
    879 
    880 5. ST_Histogram(rast raster, nband int) -> set of records
    881 
    882   parameter "hasnodata" is removed and assumed to be FALSE
    883 
    884 6. ST_Histogram(rast raster) -> set of records
    885 
    886   assumes that nband is 1.
    887 
    888 7. ST_Histogram(rast raster, nband int, bins int, width double precision[], right boolean) -> set of records
    889 
    890   hasnodata is assumed to be FALSE
    891 
    892 8. ST_Histogram(rast raster, nband int, bins int, right boolean) -> set of records
    893 
    894   all bins will have equal widths
    895 
    896 9. ST_Histogram(rast raster, nband int, bins int) -> set of records
    897 
    898   right is assumed to be FALSE
    899 
    900 ST_ApproxHistogram should have the following variations.
    901 
    902 1. ST_ApproxHistogram(rast raster, nband int, hasnodata boolean, sample_percent double precision, bins int, width double precision[], right boolean) -> set of record
    903 
    904     sample_percent: a value between 0 and 1 indicating the percentage of the raster band's pixels to consider when generating the histogram.
    905 
    906 {{{
    907 ST_Histogram(rast, 2, FALSE, 0.1, NULL, NULL, FALSE)
    908 
    909 ST_Histogram(rast, 1, TRUE, 1, 100, NULL, FALSE)
    910 
    911 ST_Histogram(rast, 2, FALSE, 0.2, NULL, ARRAY[100, 50], FALSE)
    912 
    913 ST_Histogram(rast, 2, FALSE, 0.25, 9, ARRAY[1000], TRUE)
    914 
    915 ST_Histogram(rast, 2, FALSE, 0.05, 20, ARRAY[100, 200, 300], TRUE)
    916 }}}
    917 
    918 2. ST_ApproxHistogram(rast raster, nband int, hasnodata boolean, sample_percent double precision, bins int, right boolean) -> set of records
    919 
    920   parameter "width" is not specified thus resulting in all bins having the same widths
    921 
    922 3. ST_ApproxHistogram(rast raster, nband int, hasnodata boolean, sample_percent double precision, bins int) -> set of records
    923 
    924   the parameter "right" is removed and assumed to be FALSE
    925 
    926 {{{
    927 ST_ApproxHistogram(rast, 2, FALSE, 5)
    928 }}}
    929 
    930 4. ST_ApproxHistogram(rast raster, nband int, hasnodata boolean, sample_percent double precision) -> set of records
    931 
    932   the parameter "bins" is removed and set to NULL so that function can compute the number of bins to use
    933 
    934 5. ST_ApproxHistogram(rast raster, nband int, sample_percent double precision) -> set of records
    935 
    936   parameter "hasnodata" is removed and assumed to be FALSE
    937 
    938 6. ST_ApproxHistogram(rast raster, nband int) -> set of records
    939 
    940   assumes that sample_percent is 0.1
    941 
    942 7. ST_ApproxHistogram(rast raster, sample_percent double_precision) -> set of records
    943 
    944   assumes that nband is 1
    945 
    946 8. ST_ApproxHistogram(rast raster) -> set of records
    947 
    948   assumes that nband is 1 and sample_percent is 0.1
    949 
    950 9. ST_ApproxHistogram(rast raster, nband int, sample_percent double precision, bins int, width double precision[], right boolean) -> set of records
    951 
    952   hasnodata is assumed to be FALSE
    953 
    954 10. ST_ApproxHistogram(rast raster, nband int, sample_percent double precision, bins int, right boolean) -> set of records
    955 
    956   all bins will have equal widths
    957 
    958 11. ST_ApproxHistogram(rast raster, nband int, sample_percent double precision, bins int) -> set of records
    959 
    960   right is assumed to be FALSE
    961102
    962103----
    … …  
    18811022
    18821023----
     1024
     1025'''ST_SummaryStats(raster, nband) -> record'''[[BR]]
     1026This is the core function that gets the summary statistics (# of values, mean, standard deviation, minimum value, maximum value) of a specified raster's band.  It is this function that ST_Mean, ST_StdDev and ST_MinMax calls for their appropriate values.
     1027
     10281. ST_SummaryStats(rast raster, nband int, hasnodata boolean) -> record
     1029
     1030  returns one record of five columns (count, mean, stddev, min, max)
     1031
     1032  nband: index of band
     1033
     1034  hasnodata: if FALSE, any pixel who's value is nodata is ignored
     1035
     1036{{{
     1037ST_SummaryStats(rast, 1, FALSE)
     1038}}}
     1039
     10402. ST_SummaryStats(rast raster, nband int) -> record
     1041
     1042  assumes hasnodata = FALSE
     1043
     1044{{{
     1045ST_SummaryStats(rast, 2)
     1046}}}
     1047
     10483. ST_SummaryStats(rast raster, hasnodata boolean) -> record
     1049
     1050  assumes nband = 1
     1051
     1052{{{
     1053ST_SummaryStats(rast, TRUE)
     1054}}}
     1055
     10564. ST_SummaryStats(rast raster) -> record
     1057
     1058  assumes nband = 1 and hasnodata = FALSE
     1059
     1060{{{
     1061ST_SummaryStats(rast)
     1062}}}
     1063
     1064Due to the time it may take to do on-the-fly calculation of summary stats for large rasters (say 10000 x 10000), an alternative that sacrifices accuracy for speed is required.  The following functions sample a percentage of the raster in a methodical randomized manner.  The algorithm used for sampling is...
     1065
     10661. select the larger dimension of the width and height.  compute the number of pixels to sample in each "row" of the larger dimension
     1067
     10682. pick pixels from each "row" of the larger dimension in an incremental rolling manner where each increment is randomly determined.
     1069
     1070The set of ST_ApproxSummaryStats functions are:
     1071
     10721. ST_ApproxSummaryStats(rast raster, nband int, hasnodata boolean, sample_percent double precision) -> record
     1073
     1074  sample_percent: a value between 0 and 1 indicating the percentage of the raster band's pixels to consider
     1075
     1076{{{
     1077ST_ApproxSummaryStats(rast, 3, FALSE, 0.1)
     1078
     1079ST_ApproxSummaryStats(rast, 1, TRUE, 0.5)
     1080}}}
     1081
     10822. ST_ApproxSummaryStats(rast raster, nband int, sample_percent double precision) -> record
     1083
     1084  assumes that nband = 1
     1085
     1086{{{
     1087ST_ApproxSummaryStats(rast, 2 0.01)
     1088
     1089ST_ApproxSummaryStats(rast, 4, 0.025)
     1090}}}
     1091
     10923. ST_ApproxSummaryStats(rast raster, hasnodata boolean, sample_percent double precision) -> record
     1093
     1094  assumes that nband = 1
     1095
     1096{{{
     1097ST_ApproxSummaryStats(rast, FALSE, 0.01)
     1098
     1099ST_ApproxSummaryStats(rast, TRUE, 0.025)
     1100}}}
     1101
     11024. ST_ApproxSummaryStats(rast raster, sample_percent double precision) -> record
     1103
     1104  assumes that nband = 1 and hasnodata = FALSE
     1105
     1106{{{
     1107ST_ApproxSummaryStats(rast, 0.25)
     1108}}}
     1109
     11105. ST_ApproxSummaryStats(rast raster) -> record
     1111
     1112  assumes that nband = 1, hasnodata = FALSE and sample_percent = 0.1
     1113
     1114{{{
     1115ST_ApproxSummaryStats(rast)
     1116}}}
     1117
     1118The situation arises where the summary statistics of a coverage table is required.  As the coverage may be large (tens of gigabytes of memory or larger), the following functions are provided to permit an incremental computation of the summary statistics.
     1119
     11201. ST_SummaryStats(rastertable text, rastercolumn text, nband int, hasnodata boolean) -> record
     1121
     1122  rastertable: name of table with raster column
     1123
     1124  rastercolumn: name of column of data type raster
     1125
     1126{{{
     1127ST_SummaryStats('tmax_2010', 'rast', 1, FALSE)
     1128
     1129ST_SummaryStats('precip_2011', 'rast', 1, TRUE)
     1130}}}
     1131
     11322. ST_SummaryStats(rastertable text, rastercolumn text, nband int) -> record
     1133
     1134    hasnodata = FALSE
     1135
     1136{{{
     1137ST_SummaryStats('tmax_2010', 'rast', 1)
     1138}}}
     1139
     11403. ST_SummaryStats(rastertable text, rastercolumn text, hasnodata boolean) -> record
     1141
     1142    nband = 1
     1143
     1144{{{
     1145ST_SummaryStats('precip_2011', 'rast', TRUE)
     1146}}}
     1147
     11484. ST_SummaryStats(rastertable text, rastercolumn text) -> record
     1149
     1150    nband = 1 and hasnodata = FALSE
     1151
     1152{{{
     1153ST_SummaryStats('tmin_2009', 'rast')
     1154}}}
     1155
     1156Variations for ST_ApproxSummaryStats are:
     1157
     11581. ST_ApproxSummaryStats(rastertable text, rastercolumn text, nband int, hasnodata boolean, sample_percent double precision) -> record
     1159
     1160{{{
     1161ST_ApproxSummaryStats('tmax_2010', 'rast', 1, FALSE, 0.5)
     1162
     1163ST_ApproxSummaryStats('precip_2011', 'rast', 1, TRUE, 0.2)
     1164}}}
     1165
     11662. ST_ApproxSummaryStats(rastertable text, rastercolumn text, nband int, sample_percent double precision) -> record
     1167
     1168    hasnodata = FALSE
     1169
     1170{{{
     1171ST_ApproxSummaryStats('tmax_2010', 'rast', 1, 0.5)
     1172
     1173ST_ApproxSummaryStats('precip_2011', 'rast', 1, 0.2)
     1174}}}
     1175
     11763. ST_ApproxSummaryStats(rastertable text, rastercolumn text, hasnodata boolean, sample_percent double precision) -> record
     1177
     1178    nband = 1
     1179
     1180{{{
     1181ST_ApproxSummaryStats('tmax_2010', 'rast', FALSE, 0.5)
     1182
     1183ST_ApproxSummaryStats('precip_2011', 'rast', TRUE, 0.2)
     1184}}}
     1185
     11864. ST_ApproxSummaryStats(rastertable text, rastercolumn text, sample_percent double precision) -> record
     1187
     1188    nband = 1 and hasnodata = FALSE
     1189
     1190{{{
     1191ST_ApproxSummaryStats('tmax_2010', 'rast', 0.5)
     1192
     1193ST_ApproxSummaryStats('precip_2011', 'rast', 0.2)
     1194}}}
     1195
     11965. ST_ApproxSummaryStats(rastertable text, rastercolumn text) -> record
     1197
     1198    nband = 1, hasnodata = FALSE and sample_percent = 0.1
     1199
     1200{{{
     1201ST_ApproxSummaryStats('tmax_2010', 'rast')
     1202
     1203ST_ApproxSummaryStats('precip_2011', 'rast')
     1204}}}
     1205
     1206The mean returned in the coverage functions (has rastertable and rastercolumn arguments) is a weighted mean of the means of each raster tile. The standard deviation returned is the cumulative standard deviation of all raster tiles.
     1207
     1208----
     1209
     1210'''ST_Mean(raster, nband) -> double precision'''[[BR]]
     1211This function calls ST_SummaryStats and only returns the mean from that function.
     1212
     12131. ST_Mean(rast raster, nband int, hasnodata boolean) -> double precision
     1214
     1215  returns the mean as a double precision
     1216
     1217  nband: index of band
     1218
     1219  hasnodata: if FALSE, any pixel who's value is nodata is ignored
     1220
     1221{{{
     1222ST_Mean(rast, 1, FALSE)
     1223}}}
     1224
     12252. ST_Mean(rast raster, nband int) -> double precision
     1226
     1227  assumes hasnodata = FALSE
     1228
     1229{{{
     1230ST_Mean(rast, 2)
     1231}}}
     1232
     12333. ST_Mean(rast raster, hasnodata boolean) -> double precision
     1234
     1235  assumes nband = 1
     1236
     1237{{{
     1238ST_Mean(rast, TRUE)
     1239}}}
     1240
     12414. ST_Mean(rast raster) -> double precision
     1242
     1243  assumes nband = 1 and hasnodata = FALSE
     1244
     1245{{{
     1246ST_Mean(rast)
     1247}}}
     1248
     1249The set of ST_ApproxMean functions are:
     1250
     12511. ST_ApproxMean(rast raster, nband int, hasnodata boolean, sample_percent double precision) -> double precision
     1252
     1253  sample_percent: a value between 0 and 1 indicating the percentage of the raster band's pixels to consider
     1254
     1255{{{
     1256ST_ApproxMean(rast, 3, FALSE, 0.1)
     1257
     1258ST_ApproxMean(rast, 1, TRUE, 0.5)
     1259}}}
     1260
     12612. ST_ApproxMean(rast raster, nband int, sample_percent double precision) -> double precision
     1262
     1263  assumes that nband = 1
     1264
     1265{{{
     1266ST_ApproxMean(rast, 2 0.01)
     1267
     1268ST_ApproxMean(rast, 4, 0.025)
     1269}}}
     1270
     12713. ST_ApproxMean(rast raster, hasnodata boolean, sample_percent double precision) -> double precision
     1272
     1273  assumes that nband = 1
     1274
     1275{{{
     1276ST_ApproxMean(rast, FALSE, 0.01)
     1277
     1278ST_ApproxMean(rast, TRUE, 0.025)
     1279}}}
     1280
     12814. ST_ApproxMean(rast raster, sample_percent double precision) -> double precision
     1282
     1283  assumes that nband = 1 and hasnodata = FALSE
     1284
     1285{{{
     1286ST_ApproxMean(rast, 0.25)
     1287}}}
     1288
     12895. ST_ApproxMean(rast raster) -> double precision
     1290
     1291  assumes that nband = 1, hasnodata = FALSE and sample_percent = 0.1
     1292
     1293{{{
     1294ST_ApproxMean(rast)
     1295}}}
     1296
     1297The following functions are provided for coverage tables.
     1298
     12991. ST_Mean(rastertable text, rastercolumn text, nband int, hasnodata boolean) -> double precision
     1300
     1301  rastertable: name of table with raster column
     1302
     1303  rastercolumn: name of column of data type raster
     1304
     1305{{{
     1306ST_Mean('tmax_2010', 'rast', 1, FALSE)
     1307
     1308ST_Mean('precip_2011', 'rast', 1, TRUE)
     1309}}}
     1310
     13112. ST_Mean(rastertable text, rastercolumn text, nband int) -> double precision
     1312
     1313    hasnodata = FALSE
     1314
     1315{{{
     1316ST_Mean('tmax_2010', 'rast', 1)
     1317}}}
     1318
     13193. ST_Mean(rastertable text, rastercolumn text, hasnodata boolean) -> double precision
     1320
     1321    nband = 1
     1322
     1323{{{
     1324ST_Mean('precip_2011', 'rast', TRUE)
     1325}}}
     1326
     13274. ST_Mean(rastertable text, rastercolumn text) -> double precision
     1328
     1329    nband = 1 and hasnodata = FALSE
     1330
     1331{{{
     1332ST_Mean('tmin_2009', 'rast')
     1333}}}
     1334
     1335Variations for ST_ApproxMean are:
     1336
     13371. ST_ApproxMean(rastertable text, rastercolumn text, nband int, hasnodata boolean, sample_percent double precision) -> double precision
     1338
     1339{{{
     1340ST_ApproxMean('tmax_2010', 'rast', 1, FALSE, 0.5)
     1341
     1342ST_ApproxMean('precip_2011', 'rast', 1, TRUE, 0.2)
     1343}}}
     1344
     13452. ST_ApproxMean(rastertable text, rastercolumn text, nband int, sample_percent double precision) -> double precision
     1346
     1347    hasnodata = FALSE
     1348
     1349{{{
     1350ST_ApproxMean('tmax_2010', 'rast', 1, 0.5)
     1351
     1352ST_ApproxMean('precip_2011', 'rast', 1, 0.2)
     1353}}}
     1354
     13553. ST_ApproxMean(rastertable text, rastercolumn text, hasnodata boolean, sample_percent double precision) -> double precision
     1356
     1357    nband = 1
     1358
     1359{{{
     1360ST_ApproxMean('tmax_2010', 'rast', FALSE, 0.5)
     1361
     1362ST_ApproxMean('precip_2011', 'rast', TRUE, 0.2)
     1363}}}
     1364
     13654. ST_ApproxMean(rastertable text, rastercolumn text, sample_percent double precision) -> double precision
     1366
     1367    nband = 1 and hasnodata = FALSE
     1368
     1369{{{
     1370ST_ApproxMean('tmax_2010', 'rast', 0.5)
     1371
     1372ST_ApproxMean('precip_2011', 'rast', 0.2)
     1373}}}
     1374
     13755. ST_ApproxMean(rastertable text, rastercolumn text) -> double precision
     1376
     1377    nband = 1, hasnodata = FALSE and sample_percent = 0.1
     1378
     1379{{{
     1380ST_ApproxMean('tmax_2010', 'rast')
     1381
     1382ST_ApproxMean('precip_2011', 'rast')
     1383}}}
     1384
     1385The mean returned in the coverage functions (has rastertable and rastercolumn arguments) is a weighted mean of the means of each raster tile.
     1386
     1387----
     1388
     1389'''ST_StdDev(raster, nband) -> double precision'''[[BR]]
     1390This function calls ST_SummaryStats and only returns the standard deviation from that function.
     1391
     13921. ST_StdDev(rast raster, nband int, hasnodata boolean) -> double precision
     1393
     1394  returns the standard deviation as a double precision
     1395
     1396  nband: index of band
     1397
     1398  hasnodata: if FALSE, any pixel who's value is nodata is ignored
     1399
     1400{{{
     1401ST_StdDev(rast, 1, FALSE)
     1402}}}
     1403
     14042. ST_StdDev(rast raster, nband int) -> double precision
     1405
     1406  assumes hasnodata = FALSE
     1407
     1408{{{
     1409ST_StdDev(rast, 2)
     1410}}}
     1411
     14123. ST_StdDev(rast raster, hasnodata boolean) -> double precision
     1413
     1414  assumes nband = 1
     1415
     1416{{{
     1417ST_StdDev(rast, TRUE)
     1418}}}
     1419
     14204. ST_StdDev(rast raster) -> double precision
     1421
     1422  assumes nband = 1 and hasnodata = FALSE
     1423
     1424{{{
     1425ST_StdDev(rast)
     1426}}}
     1427
     1428The set of ST_ApproxStdDev functions are:
     1429
     14301. ST_ApproxStdDev(rast raster, nband int, hasnodata boolean, sample_percent double precision) -> double precision
     1431
     1432  sample_percent: a value between 0 and 1 indicating the percentage of the raster band's pixels to consider
     1433
     1434{{{
     1435ST_ApproxStdDev(rast, 3, FALSE, 0.1)
     1436
     1437ST_ApproxStdDev(rast, 1, TRUE, 0.5)
     1438}}}
     1439
     14402. ST_ApproxStdDev(rast raster, nband int, sample_percent double precision) -> double precision
     1441
     1442  assumes that nband = 1
     1443
     1444{{{
     1445ST_ApproxStdDev(rast, 2 0.01)
     1446
     1447ST_ApproxStdDev(rast, 4, 0.025)
     1448}}}
     1449
     14503. ST_ApproxStdDev(rast raster, hasnodata boolean, sample_percent double precision) -> double precision
     1451
     1452  assumes that nband = 1
     1453
     1454{{{
     1455ST_ApproxStdDev(rast, FALSE, 0.01)
     1456
     1457ST_ApproxStdDev(rast, TRUE, 0.025)
     1458}}}
     1459
     14604. ST_ApproxStdDev(rast raster, sample_percent double precision) -> double precision
     1461
     1462  assumes that nband = 1 and hasnodata = FALSE
     1463
     1464{{{
     1465ST_ApproxStdDev(rast, 0.25)
     1466}}}
     1467
     14685. ST_ApproxStdDev(rast raster) -> double precision
     1469
     1470  assumes that nband = 1, hasnodata = FALSE and sample_percent = 0.1
     1471
     1472{{{
     1473ST_ApproxStdDev(rast)
     1474}}}
     1475
     1476The following functions are provided for coverage tables.
     1477
     14781. ST_StdDev(rastertable text, rastercolumn text, nband int, hasnodata boolean) -> double precision
     1479
     1480  rastertable: name of table with raster column
     1481
     1482  rastercolumn: name of column of data type raster
     1483
     1484{{{
     1485ST_StdDev('tmax_2010', 'rast', 1, FALSE)
     1486
     1487ST_StdDev('precip_2011', 'rast', 1, TRUE)
     1488}}}
     1489
     14902. ST_StdDev(rastertable text, rastercolumn text, nband int) -> double precision
     1491
     1492    hasnodata = FALSE
     1493
     1494{{{
     1495ST_StdDev('tmax_2010', 'rast', 1)
     1496}}}
     1497
     14983. ST_StdDev(rastertable text, rastercolumn text, hasnodata boolean) -> double precision
     1499
     1500    nband = 1
     1501
     1502{{{
     1503ST_StdDev('precip_2011', 'rast', TRUE)
     1504}}}
     1505
     15064. ST_StdDev(rastertable text, rastercolumn text) -> double precision
     1507
     1508    nband = 1 and hasnodata = FALSE
     1509
     1510{{{
     1511ST_StdDev('tmin_2009', 'rast')
     1512}}}
     1513
     1514Variations for ST_ApproxStdDev are:
     1515
     15161. ST_ApproxStdDev(rastertable text, rastercolumn text, nband int, hasnodata boolean, sample_percent double precision) -> double precision
     1517
     1518{{{
     1519ST_ApproxStdDev('tmax_2010', 'rast', 1, FALSE, 0.5)
     1520
     1521ST_ApproxStdDev('precip_2011', 'rast', 1, TRUE, 0.2)
     1522}}}
     1523
     15242. ST_ApproxStdDev(rastertable text, rastercolumn text, nband int, sample_percent double precision) -> double precision
     1525
     1526    hasnodata = FALSE
     1527
     1528{{{
     1529ST_ApproxStdDev('tmax_2010', 'rast', 1, 0.5)
     1530
     1531ST_ApproxStdDev('precip_2011', 'rast', 1, 0.2)
     1532}}}
     1533
     15343. ST_ApproxStdDev(rastertable text, rastercolumn text, hasnodata boolean, sample_percent double precision) -> double precision
     1535
     1536    nband = 1
     1537
     1538{{{
     1539ST_ApproxStdDev('tmax_2010', 'rast', FALSE, 0.5)
     1540
     1541ST_ApproxStdDev('precip_2011', 'rast', TRUE, 0.2)
     1542}}}
     1543
     15444. ST_ApproxStdDev(rastertable text, rastercolumn text, sample_percent double precision) -> double precision
     1545
     1546    nband = 1 and hasnodata = FALSE
     1547
     1548{{{
     1549ST_ApproxStdDev('tmax_2010', 'rast', 0.5)
     1550
     1551ST_ApproxStdDev('precip_2011', 'rast', 0.2)
     1552}}}
     1553
     15545. ST_ApproxStdDev(rastertable text, rastercolumn text) -> double precision
     1555
     1556    nband = 1, hasnodata = FALSE and sample_percent = 0.1
     1557
     1558{{{
     1559ST_ApproxStdDev('tmax_2010', 'rast')
     1560
     1561ST_ApproxStdDev('precip_2011', 'rast')
     1562}}}
     1563
     1564The standard deviation returned in the coverage functions (has rastertable and rastercolumn arguments) is the cumulative standard deviation of all raster tiles.
     1565
     1566----
     1567
     1568'''ST_MinMax(raster, nband) -> record'''[[BR]]
     1569This function calls ST_SummaryStats and only returns the min and max values from that function.
     1570
     15711. ST_MinMax(rast raster, nband int, hasnodata boolean) -> record
     1572
     1573  returns the record (min double precision, max double precision)
     1574
     1575  nband: index of band
     1576
     1577  hasnodata: if FALSE, any pixel who's value is nodata is ignored
     1578
     1579{{{
     1580ST_MinMax(rast, 1, FALSE)
     1581}}}
     1582
     15832. ST_MinMax(rast raster, nband int) -> record
     1584
     1585  assumes hasnodata = FALSE
     1586
     1587{{{
     1588ST_MinMax(rast, 2)
     1589}}}
     1590
     15913. ST_MinMax(rast raster, hasnodata boolean) -> record
     1592
     1593  assumes nband = 1
     1594
     1595{{{
     1596ST_MinMax(rast, TRUE)
     1597}}}
     1598
     15994. ST_MinMax(rast raster) -> record
     1600
     1601  assumes nband = 1 and hasnodata = FALSE
     1602
     1603{{{
     1604ST_MinMax(rast)
     1605}}}
     1606
     1607The set of ST_ApproxMinMax functions are:
     1608
     16091. ST_ApproxMinMax(rast raster, nband int, hasnodata boolean, sample_percent record) -> record
     1610
     1611  sample_percent: a value between 0 and 1 indicating the percentage of the raster band's pixels to consider
     1612
     1613{{{
     1614ST_ApproxMinMax(rast, 3, FALSE, 0.1)
     1615
     1616ST_ApproxMinMax(rast, 1, TRUE, 0.5)
     1617}}}
     1618
     16192. ST_ApproxMinMax(rast raster, nband int, sample_percent double precision) -> record
     1620
     1621  assumes that nband = 1
     1622
     1623{{{
     1624ST_ApproxMinMax(rast, 2 0.01)
     1625
     1626ST_ApproxMinMax(rast, 4, 0.025)
     1627}}}
     1628
     16293. ST_ApproxMinMax(rast raster, hasnodata boolean, sample_percent double precision) -> record
     1630
     1631  assumes that nband = 1
     1632
     1633{{{
     1634ST_ApproxMinMax(rast, FALSE, 0.01)
     1635
     1636ST_ApproxMinMax(rast, TRUE, 0.025)
     1637}}}
     1638
     16394. ST_ApproxMinMax(rast raster, sample_percent double precision) -> record
     1640
     1641  assumes that nband = 1 and hasnodata = FALSE
     1642
     1643{{{
     1644ST_ApproxMinMax(rast, 0.25)
     1645}}}
     1646
     16475. ST_ApproxMinMax(rast raster) -> record
     1648
     1649  assumes that nband = 1, hasnodata = FALSE and sample_percent = 0.1
     1650
     1651{{{
     1652ST_ApproxMinMax(rast)
     1653}}}
     1654
     1655The following functions are provided for coverage tables.
     1656
     16571. ST_MinMax(rastertable text, rastercolumn text, nband int, hasnodata boolean) -> record
     1658
     1659  rastertable: name of table with raster column
     1660
     1661  rastercolumn: name of column of data type raster
     1662
     1663{{{
     1664ST_MinMax('tmax_2010', 'rast', 1, FALSE)
     1665
     1666ST_MinMax('precip_2011', 'rast', 1, TRUE)
     1667}}}
     1668
     16692. ST_MinMax(rastertable text, rastercolumn text, nband int) -> record
     1670
     1671    hasnodata = FALSE
     1672
     1673{{{
     1674ST_MinMax('tmax_2010', 'rast', 1)
     1675}}}
     1676
     16773. ST_MinMax(rastertable text, rastercolumn text, hasnodata boolean) -> record
     1678
     1679    nband = 1
     1680
     1681{{{
     1682ST_MinMax('precip_2011', 'rast', TRUE)
     1683}}}
     1684
     16854. ST_MinMax(rastertable text, rastercolumn text) -> record
     1686
     1687    nband = 1 and hasnodata = FALSE
     1688
     1689{{{
     1690ST_MinMax('tmin_2009', 'rast')
     1691}}}
     1692
     1693Variations for ST_ApproxMinMax are:
     1694
     16951. ST_ApproxMinMax(rastertable text, rastercolumn text, nband int, hasnodata boolean, sample_percent double precision) -> record
     1696
     1697{{{
     1698ST_ApproxMinMax('tmax_2010', 'rast', 1, FALSE, 0.5)
     1699
     1700ST_ApproxMinMax('precip_2011', 'rast', 1, TRUE, 0.2)
     1701}}}
     1702
     17032. ST_ApproxMinMax(rastertable text, rastercolumn text, nband int, sample_percent double precision) -> record
     1704
     1705    hasnodata = FALSE
     1706
     1707{{{
     1708ST_ApproxMinMax('tmax_2010', 'rast', 1, 0.5)
     1709
     1710ST_ApproxMinMax('precip_2011', 'rast', 1, 0.2)
     1711}}}
     1712
     17133. ST_ApproxMinMax(rastertable text, rastercolumn text, hasnodata boolean, sample_percent double precision) -> record
     1714
     1715    nband = 1
     1716
     1717{{{
     1718ST_ApproxMinMax('tmax_2010', 'rast', FALSE, 0.5)
     1719
     1720ST_ApproxMinMax('precip_2011', 'rast', TRUE, 0.2)
     1721}}}
     1722
     17234. ST_ApproxMinMax(rastertable text, rastercolumn text, sample_percent double precision) -> record
     1724
     1725    nband = 1 and hasnodata = FALSE
     1726
     1727{{{
     1728ST_ApproxMinMax('tmax_2010', 'rast', 0.5)
     1729
     1730ST_ApproxMinMax('precip_2011', 'rast', 0.2)
     1731}}}
     1732
     17335. ST_ApproxMinMax(rastertable text, rastercolumn text) -> record
     1734
     1735    nband = 1, hasnodata = FALSE and sample_percent = 0.1
     1736
     1737{{{
     1738ST_ApproxMinMax('tmax_2010', 'rast')
     1739
     1740ST_ApproxMinMax('precip_2011', 'rast')
     1741}}}
     1742
     1743----
     1744
     1745'''ST_Histogram(raster, nband) -> set of records'''[[BR]]
     1746ST_Histogram and ST_ApproxHistogram provide methods to determine a raster's data distribution.
     1747
     1748The return of ST_Histogram and ST_ApproxHistogram is a set of records where each record is (min, max, count, proportion).
     1749
     1750ST_Histogram has the following variations.
     1751
     17521. ST_Histogram(rast raster, nband int, hasnodata boolean, bins int, width double precision[], right boolean) -> set of records
     1753
     1754  returns set of records of four columns (min, max, count, proportion)
     1755
     1756  nband: index of band to process on
     1757
     1758  hasnodata: if FALSE, any pixel who's value is nodata is ignored.
     1759
     1760  bins: the number of categories/bins to have in the histogram. If NULL or value less than one, the number of categories will be auto-computed using Sturges' formula if the number of values >= 30 or Square-root choice if number of values < 30.
     1761
     1762    http://en.wikipedia.org/wiki/Histogram#Mathematical_definition
     1763
     1764  width: an array indicating the width of each category/bin. If the number of bins is greater than the number of widths, the widths are repeated. Example: 9 bins, widths are [a, b, c] will have the output be [a, b, c, a, b, c, a, b, c].
     1765
     1766  right: compute the histogram from the right rather than from the left (default). This changes the criteria for evaluating a value x from [a, b) to (a, b].
     1767
     1768{{{
     1769ST_Histogram(rast, 2, FALSE, NULL, NULL, FALSE)
     1770
     1771ST_Histogram(rast, 1, TRUE, 100, NULL, FALSE)
     1772
     1773ST_Histogram(rast, 2, FALSE, NULL, ARRAY[100, 50], FALSE)
     1774
     1775ST_Histogram(rast, 2, FALSE, 9, ARRAY[1000], TRUE)
     1776
     1777ST_Histogram(rast, 2, FALSE, 20, ARRAY[100, 200, 300], TRUE)
     1778}}}
     1779
     17802. ST_Histogram(rast raster, nband int, hasnodata boolean, bins int, right boolean) -> set of records
     1781
     1782  parameter "width" is not specified thus resulting in all bins having the same widths
     1783
     1784{{{
     1785ST_Histogram(rast, 2, FALSE, 5, FALSE)
     1786}}}
     1787
     17883. ST_Histogram(rast raster, nband int, hasnodata boolean, bins int) -> set of records
     1789
     1790  the parameter "right" is removed and assumed to be FALSE
     1791
     1792{{{
     1793ST_Histogram(rast, 2, FALSE, 5)
     1794}}}
     1795
     17964. ST_Histogram(rast raster, nband int, hasnodata boolean) -> set of records
     1797
     1798  the parameter "bins" is removed and set to NULL.  The function will compute the number of bins to use
     1799
     18005. ST_Histogram(rast raster, nband int) -> set of records
     1801
     1802  parameter "hasnodata" is removed and assumed to be FALSE
     1803
     18046. ST_Histogram(rast raster) -> set of records
     1805
     1806  assumes that nband is 1.
     1807
     18087. ST_Histogram(rast raster, nband int, bins int, width double precision[], right boolean) -> set of records
     1809
     1810  hasnodata is assumed to be FALSE
     1811
     18128. ST_Histogram(rast raster, nband int, bins int, right boolean) -> set of records
     1813
     1814  all bins will have equal widths
     1815
     18169. ST_Histogram(rast raster, nband int, bins int) -> set of records
     1817
     1818  right is assumed to be FALSE
     1819
     1820ST_ApproxHistogram should have the following variations.
     1821
     18221. ST_ApproxHistogram(rast raster, nband int, hasnodata boolean, sample_percent double precision, bins int, width double precision[], right boolean) -> set of record
     1823
     1824    sample_percent: a value between 0 and 1 indicating the percentage of the raster band's pixels to consider when generating the histogram.
     1825
     1826{{{
     1827ST_Histogram(rast, 2, FALSE, 0.1, NULL, NULL, FALSE)
     1828
     1829ST_Histogram(rast, 1, TRUE, 1, 100, NULL, FALSE)
     1830
     1831ST_Histogram(rast, 2, FALSE, 0.2, NULL, ARRAY[100, 50], FALSE)
     1832
     1833ST_Histogram(rast, 2, FALSE, 0.25, 9, ARRAY[1000], TRUE)
     1834
     1835ST_Histogram(rast, 2, FALSE, 0.05, 20, ARRAY[100, 200, 300], TRUE)
     1836}}}
     1837
     18382. ST_ApproxHistogram(rast raster, nband int, hasnodata boolean, sample_percent double precision, bins int, right boolean) -> set of records
     1839
     1840  parameter "width" is not specified thus resulting in all bins having the same widths
     1841
     18423. ST_ApproxHistogram(rast raster, nband int, hasnodata boolean, sample_percent double precision, bins int) -> set of records
     1843
     1844  the parameter "right" is removed and assumed to be FALSE
     1845
     1846{{{
     1847ST_ApproxHistogram(rast, 2, FALSE, 5)
     1848}}}
     1849
     18504. ST_ApproxHistogram(rast raster, nband int, hasnodata boolean, sample_percent double precision) -> set of records
     1851
     1852  the parameter "bins" is removed and set to NULL so that function can compute the number of bins to use
     1853
     18545. ST_ApproxHistogram(rast raster, nband int, sample_percent double precision) -> set of records
     1855
     1856  parameter "hasnodata" is removed and assumed to be FALSE
     1857
     18586. ST_ApproxHistogram(rast raster, nband int) -> set of records
     1859
     1860  assumes that sample_percent is 0.1
     1861
     18627. ST_ApproxHistogram(rast raster, sample_percent double_precision) -> set of records
     1863
     1864  assumes that nband is 1
     1865
     18668. ST_ApproxHistogram(rast raster) -> set of records
     1867
     1868  assumes that nband is 1 and sample_percent is 0.1
     1869
     18709. ST_ApproxHistogram(rast raster, nband int, sample_percent double precision, bins int, width double precision[], right boolean) -> set of records
     1871
     1872  hasnodata is assumed to be FALSE
     1873
     187410. ST_ApproxHistogram(rast raster, nband int, sample_percent double precision, bins int, right boolean) -> set of records
     1875
     1876  all bins will have equal widths
     1877
     187811. ST_ApproxHistogram(rast raster, nband int, sample_percent double precision, bins int) -> set of records
     1879
     1880  right is assumed to be FALSE
     1881
     1882----
    18831883== '''Objective FV.17 - Being able to refer to band by textual name.''' ==
    18841884