I have a situation where I have a Hive table that is time partitioned by
individual fields (e.g. time_year, time_month, time_day, time_hour). I
want to regularly run Hive queries over this table for data in the last 24
hours (or X days or X months). To improve performance and make use of
these partitions to cut out data I'm trying to figure out how to calculate
and parse the components of a time value based on the nominal time.
To calculate the time I want I know I can do the following:
${coord:dateOffset(coord:nominalTime(), -1, 'DAY')}
So now what I'm trying to figure out is how to parse the individual year,
month, day, hour components out of that calculate time similar to how you
can have the ${YEAR}, ${MONTH}, etc values when composing a dataset. Then
I'll be able to substitute those values into a Hive query like
>> select * from table foo with time_year >= ${CALCULATED_YEAR}...;
Can this be done using existing EL Functions? I'd prefer not to have to
write my own.
Micah
