[SPARK-12014] Spark SQL query containing semicolon is broken in Beeline (related to HIVE-11100)

• Type: Bug
• Status: Resolved
• Priority: Minor
• Resolution: Incomplete
• Affects Version/s: 1.5.2
• Component/s: SQL
• Labels: bulk-closed
• Assignee: Unassigned
• Reporter: Teng Qiu
• Created: 26/Nov/15 13:30
• Updated: 12/Oct/19 02:38
• Resolved: 08/Oct/19 05:42

Description

Actually it is known hive issue: https://issues.apache.org/jira/browse/HIVE-11100

patch available: https://reviews.apache.org/r/35907/diff/1

but Spark uses its own maven dependencies for hive (org.spark-project.hive), we can not use this patch to fix the problem, it would be better if you can fix this in spark's hive package.

In spark's beeline, the error message will be:

0: jdbc:hive2://host:10000/> CREATE TABLE beeline_tb (c1 int, c2 string) ROW FORMAT DELIMITED FIELDS TERMINATED BY ';' LINES TERMINATED BY '\n';
Error: org.apache.spark.sql.AnalysisException: mismatched input '<EOF>' expecting StringLiteral near 'BY' in table row format's field separator; line 1 pos 87 (state=,code=0)

0: jdbc:hive2://host:10000/> CREATE TABLE beeline_tb (c1 int, c2 string) ROW FORMAT DELIMITED FIELDS TERMINATED BY '\;' LINES TERMINATED BY '\n';
Error: org.apache.spark.sql.AnalysisException: mismatched input '<EOF>' expecting StringLiteral near 'BY' in table row format's field separator; line 1 pos 88 (state=,code=0)

0: jdbc:hive2://host:10000/> SELECT str_to_map(other_data,';','=')['key_name'] FROM some_logs WHERE log_date = '20151125' limit 5;
Error: org.apache.spark.sql.AnalysisException: cannot recognize input near '<EOF>' '<EOF>' '<EOF>' in select expression; line 1 pos 30 (state=,code=0)

0: jdbc:hive2://host:10000/> SELECT str_to_map(other_data,'\;','=')['key_name'] FROM some_logs WHERE log_date = '20151125' limit 5;
Error: org.apache.spark.sql.AnalysisException: cannot recognize input near '<EOF>' '<EOF>' '<EOF>' in select expression; line 1 pos 31 (state=,code=0)

Issue Links

• is duplicated by: SPARK-12497 thriftServer does not support semicolon in sql
• relates to: HIVE-11100 Beeline should escape semi-colon in queries
• relates to: SPARK-23710 Upgrade the built-in Hive to 2.3.5 for hadoop-3.2