[Feature Request] Better keywords control

[kavimaluskam]

Problem

Common keywords in SQL query like uuid, private, version are identified as keywords and affected tokenization result.

Proposal

If user can choose the type of SQL engine they are using, e.g.

tokens = sqlparse.parse(sql, engine='postgresql')[0]

so is_keyword function will only use common_keywords and psql_keywords during the parsing

[silpheel]

Similarly, identifiers like box which are not reserved in MySQL but are e.g. in PostgreSQL get capitalized, so I'd also like some way to specify the engine if possible. Meanwhile, back ticks can be used to work around this.

[hychen20]

Similar issue here, final is recognized as a keyword, but it shouldn't be given it is in MySQL. For example:

import sqlparse
sql = """
with final as (
    select * from test
) select * from final;
"""
stmt = sqlparse.parse(sql)[0]
print(stmt.tokens)

Output:

[<Newline ' ' at 0x7FB98003BAC0>, <CTE 'with' at 0x7FB98019EDC0>, <Whitespace ' ' at 0x7FB98019EE20>, <Keyword 'final' at 0x7FB98019EE80>, <Whitespace ' ' at 0x7FB98019EEE0>, <Keyword 'as' at 0x7FB98019EF40>, <Whitespace ' ' at 0x7FB98019EFA0>, <Parenthesis '( ...' at 0x7FB980160F90>, <Whitespace ' ' at 0x7FB9801A65E0>, <DML 'select' at 0x7FB9801A6640>, <Whitespace ' ' at 0x7FB9801A66A0>, <Wildcard '*' at 0x7FB9801A6700>, <Whitespace ' ' at 0x7FB9801A6760>, <Keyword 'from' at 0x7FB9801A67C0>, <Whitespace ' ' at 0x7FB9801A6820>, <Keyword 'final' at 0x7FB9801A6880>, <Punctuation ';' at 0x7FB9801A68E0>]