Re: unable to drop index because it does not exists

>
> Partition key: LIST (date_part('year'::text, mis_ora))
>

As an aside, you may benefit from switching to range partitioning*
depending on how your queries are written. If you have conditions such as
"WHERE mis_ora BETWEEN CURRENT_DATE - 30 AND CURRENT_DATE" or similar, then
the fact that your partitioning is defined as a function result will mean
all partitions gets scanned instead of partitioned being pruned as early as
possible in the process. That's my understanding anyway. If you always
include date_part( 'year', mis_ora) comparison in your where/join
conditions, then you'll likely be just fine. Do as you need.

*eg '01/01/2018' to '01/01/2019' for the 2018 partition since upper bound
is always exclusive