无法从Hive表中检索正确的最大日期？

Question

I am abit confused. When I use max(send_date) I get 2018-02-04 23:59:51.0 . However, when I use max(TO_DATE(from_unixtime(UNIX_TIMESTAMP(send_date, 'yyyy-mm-dd')))) or TO_DATE(from_unixtime(max(UNIX_TIMESTAMP(send_date, 'yyyy-mm-dd')))) , I get 2018-01-31 !!

Why?

select max(send_date) from mytable;

Aso, when I use WHERE TO_DATE(from_unixtime(UNIX_TIMESTAMP(send_date, 'yyyy-mm-dd'))) = '2018-02-04' , I get the restult 0 , but it's not true.

Answer 1

I was using incorrect conversion format. In particular, I used yyyy-mm-dd instead of yyyy-MM-dd

More details can be found here: http://bigdataprogrammers.com/string-date-conversion-hive/

Answer 2

When using to_date with Hive you don't even have to specify that format because by default it returns that format. Here is an example

  select max(to_date(d1)), max(d1), min(to_date(d1)), min(d1) from (
    select '2018-02-04 23:59:51.0' as d1
    union all
    select '2018-02-04 23:59:59.0' as d1
    union all
    select '2018-01-31 23:59:51.0' as d1
    union all
    select '2018-01-31 23:59:59.0' as d1
  ) tbl 

Output

OK
2018-02-04  2018-02-04 23:59:59.0   2018-01-31  2018-01-31 23:59:51.0
Time taken: 27.547 seconds, Fetched: 1 row(s)