CODEWITHSUNDEEP
pythonpandas
RedRaven
RedRaven

Reputation: 735

Best way to create a dataframe from several lists

I am working through ThinkStats, but decided to learn Pandas along the ways as well. So the code below reads in data from a file, does some checking and then appends the data to a list. I end up with several lists containing the data I need. The code below works (except for scrambling up the columns...)

My question is: What is the best way to build a dataframe from these lists? More generally, am I accomplishing my goal in the most efficient manner? 

preglength = []
caseid = []
outcome = []
birthorder = []
finalweight = []

with open('2002FemPreg.dat') as f:
    for line in f:
            caseid.append(int(line[0:13].strip()))
            preglength.append(int(line[274:276].strip()))
            outcome.append(int(line[276].strip()))
            try:
                    birthorder.append(int(line[277:279]))
            except ValueError:
                    birthorder.append(np.nan)
            finalweight.append(float(line[422:440].strip()))

c1 = pd.Series(caseid)
c2 = pd.Series(preglength)
c3 = pd.Series(outcome)
c4 = pd.Series(birthorder)
c5 = pd.Series(finalweight)


data = pd.DataFrame({'caseid': c1,'preglength': c2,'outcome': c3,'birthorder':    c4,'weight': c5})

print(data.head())

Upvotes: 0

Views: 108

Answers (1)

DSM
DSM

Reputation: 353419

I would probably use read_fwf:

>>> df = pd.read_fwf("2002FemPreg.dat",
... colspecs=[(0,13), (274, 276), (276, 277), (277, 279), (422, 440)],
... names=["caseid", "preglength", "outcome", "birthorder", "finalweight"])
>>> df.head()
   caseid  preglength  outcome  birthorder   finalweight
0       1          39        1           1   6448.271112
1       1          39        1           2   6448.271112
2       2          39        1           1  12999.542264
3       2          39        1           2  12999.542264
4       2          39        1           3  12999.542264

[5 rows x 5 columns]

Upvotes: 3

Related Questions