Hi Stamatis,
Thanks for the quick response, I tried your suggestion as follows
RelToSqlConverter converter = new RelToSqlConverter(new
AnsiSqlDialect(AnsiSqlDialect.DEFAULT_CONTEXT.withIdentifierQuoteString("")));
SqlSelect select = converter.visitRoot(relNode).asSelect();
SqlWriterConfig config = SqlPrettyWriter.config().
withCaseClausesOnNewLines(false).
withQuoteAllIdentifiers(false);
System.out.println(new SqlPrettyWriter(config).format(select));
and I still get this
SELECT "a"
FROM test_table
WHERE "b" = 'value';
It is not obvious to me as to what I am doing wrong. Is the above block of
code the only pieces that I need to put in play to get that sql generated?
--Sandeep
On Thu, May 26, 2022 at 2:22 PM Stamatis Zampetakis <[email protected]>
wrote:
> Hi Sandeep,
>
> If you want to turn off quoting (which I am not sure if it is a good idea)
> when you go from relational algebra to SQL you have to customize the
> SqlDialect that you are using via Context#withIdentifierQuoteString [1].
> You can find a small example as part of RelToSqlConverterTest here [2].
>
> Best,
> Stamatis
>
> [1]
>
> https://github.com/apache/calcite/blob/9bdfd9a178f493b235d8785afd94fd0c998e8cce/core/src/main/java/org/apache/calcite/sql/SqlDialect.java#L1408
> [2]
>
> https://github.com/zabetak/calcite/commit/ea18d28f4fb44113c414c393fe28fc94b3eecc29
>
> On Thu, May 26, 2022 at 9:18 PM Sandeep N <[email protected]>
> wrote:
>
> > Hi all,
> >
> > I am pretty aware that this question may have been asked before as I have
> > found references on stackoverflow but have yet to stumble on a response
> > which lays out the path on solving this.
> >
> > Here is the problem I am facing -
> > I am using relational algebra to generate SQL on a table test_table with
> > two columns a and b fed into the calcite Framework (see below) via a
> custom
> > catalog. Both a and b are varchar columns.
> >
> > FrameworkConfig config = Frameworks.newConfigBuilder().
> > parserConfig(SqlParser.Config.DEFAULT).
> > defaultSchema(addSchemas(rootSchema, schemaName,
> schema)).
> > <--- this is where I plug my custom table.
> > traitDefs((List<RelTraitDef>) null).
> > operatorTable(sqlOperatorTable).
> > programs(Programs.heuristicJoinOrder(Programs.RULE_SET,
> > true, 2)).build()
> >
> > The relational algebra is then used to generate SQL to the AnsiSqlDialect
> > and the resultant output SQL is
> >
> > SELECT "a"
> > FROM test_table
> > WHERE "b" = 'value';
> >
> > I have the SqlWriterConfig setup as such
> >
> > SqlWriterConfig config = SqlPrettyWriter.config().
> > withCaseClausesOnNewLines(false).
> > withQuoteAllIdentifiers(false);
> >
> > I have been trying to figure what needs to change in the sql writer but
> > after scanning stackoverflow it appears this is controlled by
> > CalciteConnectionConfig and potentially not the SqlWriter. Is this
> > accurate? and if so how do I assemble the right connection config when I
> am
> > building the custom catalog?
> >
> > P.s: I am not parsing SQL but writing directly to the relational algebra
> > nodes so SqlParser does not come into play here.
> >
> > Any pointers on how I can go about this is appreciated, say existing
> tests
> > or code which will help me put together a configuration which gets me to
> > this sql
> >
> > SELECT a
> > FROM test_table
> > WHERE b = 'value';
> >
> > Thanks in advance.
> >
> > --Sandeep
> >
>
