CODEWITHSUNDEEP
python
jlewis
jlewis

Reputation: 183

Turn a larger Pandas data frame into smaller rolling data frames

I'm new to Python, know just enough R to get by. I have a 10 by 10 dataframe.

small2
       USLC      USSC      INTD  ...      DSTS      PCAP       PRE
0  0.059304  0.019987 -0.034140  ...  0.003009  0.113144 -0.021656
1  0.003835 -0.024248  0.012446  ...  0.005323 -0.013716  0.011109
2 -0.045045 -0.047186 -0.002372  ... -0.011956 -0.118342 -0.045023
3  0.054108  0.002787  0.003714  ...  0.014466  0.128931 -0.007596
4  0.064045  0.111250  0.077478  ...  0.012059  0.115427  0.079145
5  0.041442  0.042858  0.047701  ...  0.009984  0.047098  0.003579
6  0.081832  0.046531  0.010531  ...  0.031772  0.126552  0.001398
7 -0.047171  0.022883 -0.065095  ... -0.010224 -0.025990 -0.055431
8  0.054844  0.073193  0.044514  ...  0.016301  0.031755  0.044597
9 -0.032403 -0.043930 -0.065013  ...  0.011944 -0.032902 -0.117689

I want to create a list of several dataframes that are each just rolling 5 by 10 frames. Rows 0 through 4, 1 through 5, etc. I've seen articles addressing something similar, but they haven't worked. I'm thinking about it like lapply in R.

I've tried splits = [small2.iloc[[i-4:i]] for i in small2.index] and got a syntax error from the colon.

I then tried splits = [small2.iloc[[i-4,i]] for i in small2.index] which gave me a list of ten elements. It should be six 5 by 10 elements.

Feel like I'm missing something basic. Thank you!

Upvotes: 1

Views: 47

Answers (2)

Cameron Riddell
Cameron Riddell

Reputation: 13457

You can also use the DataFrame.rolling method for this and simply toss away the first few iterations until the window reaches its full size

splits = [frame for frame in small2.rolling(5) if len(frame) == 5]

Upvotes: 1

jlewis
jlewis

Reputation: 183

I figured it out. splits = [small2.iloc[i-4:i+1] for i in small2.index[4:10]]

Not sure how this indexing makes sense though.

Upvotes: 1

Related Questions