[PATCH] OGR SQL 'SUBSTR' is not conformant with other SQL implementations

[Even Rouault]

Our implementation considers that the character offset is like C language, that is to say that the first character is offset 0, the second one is offset 1, etc... All other implementations I've found consider 1 as the first character, 2 as the second one (and as a special case, if 0 is provided, they turn it to 1...)

References:

• ​http://www.sqlite.org/lang_corefunc.html (SQLite)
• ​http://www.1keydata.com/sql/sql-substring.html (several implementations)
• ​http://www.adp-gmbh.ch/ora/sql/substr.html (Oracle)
• ​http://www.postgresql.org/docs/8.1/static/functions-string.html (PostgreSQL)
• ​http://dev.mysql.com/doc/refman/5.0/en/string-functions.html#function_substr (mysql)

Should we fix our implementation ? (SUBSTR has been introduced in GDAL 1.8.0 and is not yet advertized in ogr_sql.dox)