Use varied parameters as index?
Currently, in our processed data, varied parameters are in the dataframe as a column. Would it perhaps make more sense to use them as index instead (using e.g. DataFrame.set_index()
)?
Using them as index makes them very well suited for lookups (looking up with indices in a dataframe is done through hashing; looking up a specific value of the columns is done through much-slower filtering).
For more information, see also this stackexchange post.
This should have low priority but might be good to think about, especially if it turns out that such lookups are taking a lot of time somewhere.
Edited by Guus