no proper support for escaped symbols

[zhgzhg]

INSERT INTO ..... VALUES('TEST', 'T's Use Rate(%) (%)' , 'TT's Use Rate(%) (%)' , 'OTHER')
won't format properly.

[Shoelace]

that is not valid oracle SQL atleats.. to escape quotes in oralc ethey get doubled..

ie INSERT INTO ..... VALUES('TEST', 'T''s Use Rate(%) (%)' , 'TT''s Use Rate(%) (%)' , 'OTHER')

i dont knwo about other SQL varients

[zhgzhg]

Example: Table 9.1
http://dev.mysql.com/doc/refman/5.0/en/string-literals.html

[dnsmkl]

Hi zhgzhg,

Thanks for reporting the issue.
But...

In short:
Sorry, I can not promise you a quick resolution of this issue.
(honest proposal: if this is real problem, please use something else.
Search for "free sql formatter" gives some quite good online tools)

Longer version:
mysql syntax is incompatible with other sql syntaxes.
(i.e. if mysql case is naively fixed, then other flavors of sql syntax get broken)
Effective fix would include some mode switching to handle only single syntax at one time
("option: use mysql syntax", "option: use postgre syntax", "option: use oracle syntax")
But this is problematic because,
it requires quite significant changes to "architecture" of fsqlf (flex regex for string is hardcoded).

That said. I will definitely try to fix this. (but first I have to come up with good plan how to go about it. Proposals are welcome)

[rohitn]

Hi @dnsmkl,

For open source databases, perhaps the best solution is to use their own parser. See this interesting post

Not suggesting that fsqlf go down this path. I don't think it is worth the effort. When I have run into similar issues with SQL Server syntax oddities I modify the query so fsqlf can format it and then undo the modifications.

Unrelated news. I have command line fsqlf building and running on OSX. There are some odd differences in the output compared to Ubuntu. Will submit a PR after I track those down.

[dnsmkl]

Hi @rohitn,

thanks for the link and idea.
You are completely right, best solution would be to use parser of actual database.
(or at least parser that is recognized as very good for that db syntax)
..but I don't have that much energy&time that this would require.