Ryan Blue created SPARK-16178:
---------------------------------
Summary: SQL - Hive writer should not require partition names to
match table partitions
Key: SPARK-16178
URL: https://issues.apache.org/jira/browse/SPARK-16178
Project: Spark
Issue Type: Sub-task
Reporter: Ryan Blue
SPARK-14459 added a check that the {{partition}} metadata on
{{InsertIntoTable}} must match the table's partition column names. But if
{{partitionBy}} is used to set up partition columns, those columns may not be
named or the names may not match.
For example:
{code}
// Tables:
// CREATE TABLE src (id string, date int, hour int, timestamp bigint);
// CREATE TABLE dest (id string, timestamp bigint, c1 string, c2 int)
// PARTITIONED BY (utc_dateint int, utc_hour int);
spark.table("src").write.partitionBy("date", "hour").insertInto("dest")
{code}
The call to partitionBy correctly places the date and hour columns at the end
of the logical plan, but the names don't match the "utc_" prefix and the write
fails. But the analyzer will verify the types and insert an {{Alias}} so the
query is actually valid.
--
This message was sent by Atlassian JIRA
(v6.3.4#6332)
---------------------------------------------------------------------
To unsubscribe, e-mail: issues-unsubscr...@spark.apache.org
For additional commands, e-mail: issues-h...@spark.apache.org
