CODEWITHSUNDEEP
pythonpython-3.xpandasdataframe
olo
olo

Reputation: 163

How to group timestamp and count items according to days interval?

The data is like:(Note that the dates are not consecutive)

                name created    label         
0   leahbirdjohnso2 2020-02-20  PATRIOTSAWAKENED    
1   leahbirdjohnso2 2020-02-21  TRUMP2020           
2   carol2busy      2020-02-23  TRUMP2020   
3   carol2busy      2020-02-24  TRUMP2020   
4   GODRUS          2020-02-25  FOXNEWS

if I set interval =2 days,then I got a dataframe:

created     counts  label
2020-02-20  1   PATRIOTSAWAKENED
            1   TRUMP2020
2020-02-23  1   TRUMP2020       
2020-02-24  1   FOXNEWS
            1   TRUMP2020

if I set interval = 3 days,then I got a dataframe:

    created     counts  label
    2020-02-20  1   PATRIOTSAWAKENED
                1   TRUMP2020
    2020-02-23  2   TRUMP2020
                1   FOXNEWS

Basically, I want to sum up label numbers according to label names based on time intervals. I can set any intervals by days, 3days, 7days, 15days,etc. I cheked https://stackoverflow.com/questions/56275425/pandas-typeerror-not-supported-between-instances-of-int-and-str-when-s; but not working for me. I can use Counter to do it in a more complex way. How to do it in an elegant pandas style?

Upvotes: 0

Views: 350

Answers (1)

Nk03
Nk03

Reputation: 14949

IIUC, you can try:

df.created  = pd.to_datetime(df.created)
df = df.groupby([pd.Grouper(key="created", freq="3D"), 'label']).count()

Upvotes: 1

Related Questions