Is there a function in Python3 to copy the 3 smallest values in a row into a data frame, for m rows?

Question

I'm trying to find the three smallest values for each row of a dataframe, and put them in a separate dataframe. I don't need to know which column they came from, but I do need to cycle through m rows where m might change for each dataframe I use.

I wanted to use heaps.nsmallest, but I'm not sure how to loop through each row and add the results to a new line of a dataframe each time. I seem to just get a single line of results as output.

 for x in range(len(df1)):
        heap=pd.DataFrame(heapq.nsmallest(3, df1[x]))

I expected this to loop through values of x, but it only produces one column with len(df1) rows. I think it's overwriting the previous results, as it always gives the three minimum values from the last row.

Yes, you're right. I've tried creating heap[x] and it works! — Tom, May 26 '19 at 16:34
For some reason, I've reopened the notebook, and now it says "Cannot set a frame with no defined index and a value that cannot be converted to a Series" — Tom, May 26 '19 at 17:03

score 1 · Accepted Answer · answered May 26 '19 at 22:56

1

df2 = pd.DataFrame([heapq.nsmallest(3, df1[x])
                   for x in range(len(df1))])

or, use sorted(df1[x])[:3]

answered May 26 '19 at 22:56

J_H

17,926
4
24
44

Is there a function in Python3 to copy the 3 smallest values in a row into a data frame, for m rows?

1 Answers1