big-query data partitioning based on column name

0
votes

I have data-set like this 

Id  type     entry-date             userid   dashboard

  1    0   2017-12-14 00:00:08       123        A
  2    1   2017-12-15 00:00:08       234        A
  3    1   2018-04-14 00:00:08       345        A
  4    0   2017-12-14 00:00:08       456        B
  5    0   2017-12-14 00:00:08       567        B

My requirement is to

  1. Partitioned the data based on dashboard column.Is it possible in BQ to partitioned based on Column?

  2. Partitioned the data based on entrydate. I already have 600TB of data. I want to Create partitioned table based on entry date and dump existing data into this partition table. Is it possible?

  3. If 1 and 2 are not possible, what are alternatives to it?

1
databasegoogle-bigquerypartitioning
No, you can only partition by a date or timestamp column, like entry_date. - Elliott Brossard
I need data based on dashboard, if I partitioned based on entry_date, I need to query using "where dashboard=A" clause, which means scanning all the data. - user1115163
If there are not that many dashboard types, you can create separate date-partitioned tables for each one, where the dashboard type appears in the table name, then use a wildcard table to query them. - Elliott Brossard
@ElliottBrossard can you add your comments as an answer? - VictorGGl

1 Answers

0
votes

No, you can only partition by a date or timestamp column, like entry_date.

If you need data based on dashboard and there are not that many dashboard types, you can create separate date-partitioned tables for each one, where the dashboard type appears in the table name, then use a wildcard table to query them.