On BigQuery we default to 'D' daily partitions. That doesn't really make much sense. We should default to 'M' instead.
We could also think about making this adaptive. If the frontend RDBMS is partitioned on the same column then why not look at the gaps between high values and estimate whether to use 'D', 'M' or 'Y'.
On BigQuery we default to 'D' daily partitions. That doesn't really make much sense. We should default to 'M' instead.
We could also think about making this adaptive. If the frontend RDBMS is partitioned on the same column then why not look at the gaps between high values and estimate whether to use 'D', 'M' or 'Y'.