Hi :
link_crtd_date is a string of format "-MM-dd" not timestamp.
select link_crtd_date from bsl12.email_edge_lyh_mth1 limit 10;
2018-01-01
2018-01-01
2018-01-01
2018-01-01
2018-01-01
2018-01-01
2018-01-01
2018-01-01
2018-01-01
2018-01-01
Best Regards
Kelly Zhang
It seems that column `link_crtd_date` is of type `timestamp` and you therefore
partition by date including time, which produces a huge amount of directories.
I assume your intent is to partition by date (partition_date=-MM-dd or
year=/month=MM/day=dd) so you need to format/split your tim
Hi all:
when i use spark dynamic partition feature , i met a problem about hdfs
quota. I found that it is every easy to meet quota problem (exceed the max
value of quota of directory)
I have generated a unpartitioned table 'bsl12.email_edge_lyh_mth1' which
contains 584M records and will ins
