如何删除 Hive 默认分区值，__HIVE_DEFAULT_PARTITION__

Question

I am using Glue metestore for Hive and I have AWS EMR cluster to query and modify Hive tables.

My data for this resides in S3

I have 3 partition columns:

yr_no: int, month_no: int, uniq_id: int

However I have some files that have NULL values for these which Hive sets the partition value for yr_no and month_no as HIVE_DEFAULT_PARTITION

But since yr_no and month_no are both ints, I cannot easily drop those partitions.

I tried below formats:

ALTER TABLE table DROP PARTITION(yr_no=__HIVE_DEFAULT_PARTITION__);
ALTER TABLE table DROP PARTITION(yr_no<1);

First command complained since the column is int and 2nd complained about the syntax <

Is there a simple way to drop it on yr_no= HIVE_DEFAULT_PARTITION or month_no= HIVE_DEFAULT_PARTITION

Answer 1

Maybe you can use show partitons at first. And then, use desc formatted partition(yr_no=xxx,month_no=xxxx) to see the detail informations.

Answer 2

您可以执行以下操作，它会起作用。

ALTER TABLE table DROP PARTITION(yr_no='__HIVE_DEFAULT_PARTITION__');