CODEWITHSUNDEEP
pythonpandas
Christian Sauer
Christian Sauer

Reputation: 10909

Pandas groupby columns without multiindex

I have a dataframe with a data about train stations in a month, of which three are indexes: Station, Date, Hour. I could look like this:

Station    Date       Hour    Passengers 
Berlin HBF 2012-12-24 12:00   1000 
Berlin HBF 2012-12-24 13:00   2000  
Berlin HBF 2012-12-24 14:00   1000  
Berlin HBF 2012-12-24 15:00   1000  
....
Stuttgart 2012-12-24 12:00    500

Since I am only interested in sums for a station in a month, I would like to groupby by Station, Date, and Hour, so that the end result looks like this:

Station    Passengers 
Berlin HBF 4000 
....
Stuttgart  500

But I am unable to get pandas to this solution, I tried: byStation = traindata.groupby(['Station', 'Date', 'Hour']).agg(np.sum()) But that simply returns a multiindex, with all rows...

Upvotes: 2

Views: 1154

Answers (1)

dorvak
dorvak

Reputation: 9709

Looks like you want to group by "Station" only and do a sum over the "Passangers"-column. You do not need a multi-index here. Your solution will create one, but as it is the same one as your raw data, it's quite useless.

This one should work:

traindata.groupby("Station").Passengers.sum()

Upvotes: 2

Related Questions