Is merge on date and time columns faster than merge on datetime?

[PDG_VL]

Hi all,

I've been working with datatime ranges recently, e.g.: I have to exclude parts of the data based on datetime.

The process takes ages and I was wondering if PQ were faster if I split the datetime columns into date and time. So that I can merge first on date and then on time (same step, but merging on multiple columns).

My logic behind this is as follows: the number of unique values is way less in case of dates then in case of datetimes. So, for Power Query to find/filter the relevant rows from the to-be-merged table is faster. Once PQ has the result based on date, it's also faster/easier to filter further and get the relevant records based on time.

Frankly, I'm not exactly familiar how PQ performs the merge step, so if someone has a good explanation, article/video, I'd be happy to use that as a source to lear more.

Thanks,

PDG

[Anonymous · Accepted solution]

Depends on cardinality, but yes it should be much faster.

--Nate