You signed in with another tab or window. Reload to refresh your session.You signed out in another tab or window. Reload to refresh your session.You switched accounts on another tab or window. Reload to refresh your session.Dismiss alert
In #5909, we've cleanly separated the two features:
Quoting of names
Upper / lower casing of names
No new features were introduced, just the separation was implemented. However, the behaviour of RenderNameCase.LOWER and UPPER is not so terribly useful when we specify RenderQuotedNames.ALWAYS, as it might lead to wrong SQL (such as referencing "INFORMATION_SCHEMA" on PostgreSQL, or "information_schema" on H2).
This is why we need additional LOWER_IF_UNQUOTED and UPPER_IF_UNQUOTED flags in RenderNameCase, which modify names only if they're not quoted. (i.e. DSL.unquotedName()). The following test shows the expected behaviour:
In #5909, we've cleanly separated the two features:
No new features were introduced, just the separation was implemented. However, the behaviour of
RenderNameCase.LOWER
andUPPER
is not so terribly useful when we specifyRenderQuotedNames.ALWAYS
, as it might lead to wrong SQL (such as referencing"INFORMATION_SCHEMA"
on PostgreSQL, or"information_schema"
on H2).This is why we need additional
LOWER_IF_UNQUOTED
andUPPER_IF_UNQUOTED
flags inRenderNameCase
, which modify names only if they're not quoted. (i.e.DSL.unquotedName()
). The following test shows the expected behaviour:The text was updated successfully, but these errors were encountered: