[CALCITE-4802] Babel parser doesn't parse IF(condition, then, else) statements

• Type: Bug
• Status: Closed
• Priority: Minor
• Resolution: Fixed
• Fix Version/s: 1.32.0
• Component/s: babel
• Labels: pull-request-available
• Assignee: Jiajun Xie
• Reporter: Florent Martineau
• Created: 25/Sep/21 18:17
• Updated: 11/Aug/22 02:24
• Resolved: 09/Aug/22 02:07

Description

Following my email on Calcite's dev mailing list, julianhyde told me to open an issue as there is likely a bug in Babel's parser regarding IF statements.

 

You can find the email with this title: Error parsing DATE("2021-01-01") for BigQuery using Calcite

 

The symptoms are the following:

• If I use SqlParserImpl.FACTORY, most statements work but DATE(string) fails, likely because it's a reserved keyword, as Julian pointed out in the mail thread. Statements such as IF(cond, then, else) work, as well as other dialect-specific functions such as CURRENT_DATETIME() [Bigquery].
• If I use SqlBabelParserImpl.FACTORY, it will work for parsing DATE(string), it also works for dialect-specific functions, but it doesn't work for IF(cond, then, else).

 

Here is a table to sum it up:

 

 

Test	Normal parser	Babel
1+(2*4) as foo	SUCCESS	SUCCESS
CURRENT_DATE()	SUCCESS	SUCCESS
CURRENT_DATETIME()	SUCCESS	SUCCESS
DATE_FROM_UNIX(123456)	SUCCESS	SUCCESS
IF(TRUE, 1, 2)	SUCCESS	FAILURE
DATE("2021-01-01")	FAILURE	SUCCESS

 

 

Here is the code I use to check:

 

 

package test.Test;

import org.apache.calcite.config.Lex;
import org.apache.calcite.sql.SqlNode;
import org.apache.calcite.sql.parser.SqlParseException;
import org.apache.calcite.sql.parser.SqlParser;
import org.apache.calcite.sql.parser.babel.SqlBabelParserImpl;
import org.apache.calcite.sql.parser.impl.SqlParserImpl;
import org.apache.calcite.sql.validate.SqlConformanceEnum;
import org.apache.calcite.tools.FrameworkConfig;
import org.apache.calcite.tools.Frameworks;
import org.apache.calcite.tools.Planner;

import java.util.HashMap;
import java.util.Map;

public class Test {
    public static void main(String[] args) {
        SqlParser.Config sqlParserConfigBabel = SqlParser.config().DEFAULT
                .withLex(Lex.BIG_QUERY)
                .withConformance(SqlConformanceEnum.BIG_QUERY)
                .withParserFactory(SqlBabelParserImpl.FACTORY);

        SqlParser.Config sqlParserConfigNotBabel = SqlParser.config().DEFAULT
                .withLex(Lex.BIG_QUERY)
                .withConformance(SqlConformanceEnum.BIG_QUERY)
                .withParserFactory(SqlParserImpl.FACTORY);

        Map<String, SqlParser.Config> parserConfigs = new HashMap<String, SqlParser.Config>();
        parserConfigs.put("With babel", sqlParserConfigBabel);
        parserConfigs.put("Without babel", sqlParserConfigNotBabel);

        String[] testSqlFragments = {
                "1+(2*4) as foo",
                "CURRENT_DATE()",
                "CURRENT_DATETIME()",
                "DATE_FROM_UNIX(123456)",
                "IF(TRUE, 1, 2)",
                "DATE('2021-01-01')",
        };

        for (String sqlFragment : testSqlFragments) {
            String query = "SELECT " + sqlFragment;
            System.out.println("Trying to parse : " + query);

            for (var config : parserConfigs.entrySet()) {
                System.out.print(config.getKey() + " : ");
                try {
                    FrameworkConfig frameworkConfig = Frameworks.newConfigBuilder()
                            .parserConfig(config.getValue())
                            .build();
                    Planner planner = Frameworks.getPlanner(frameworkConfig);
                    SqlNode node = planner.parse(query);
                    System.out.println("SUCCESS");
                } catch (SqlParseException e) {
                    System.out.println("ERROR");
                }
            }
        }
    }
}

 

 

And here is the output of that code:

 

Trying to parse : SELECT 1+(2*4) as foo
Without babel : SUCCESS
With babel : SUCCESS
Trying to parse : SELECT CURRENT_DATE()
Without babel : SUCCESS
With babel : SUCCESS
Trying to parse : SELECT CURRENT_DATETIME()
Without babel : SUCCESS
With babel : SUCCESS
Trying to parse : SELECT DATE_FROM_UNIX(123456)
Without babel : SUCCESS
With babel : SUCCESS
Trying to parse : SELECT IF(TRUE, 1, 2)
Without babel : SUCCESS
With babel : ERROR
Trying to parse : SELECT DATE('2021-01-01')
Without babel : ERROR
With babel : SUCCESS

Process finished with exit code 0

 

 

Please tell me if I can be of any help on this issue, or if it's too deep into the project and a newcomer can't solve this.

 

Have a great day

Issue Links

• links to: GitHub Pull Request #2835