Python & Pandas: Pendant for R/dplyr .rds format -

what's closest pendant / alternative .rds format known r python respectively pandas dataframe?

i tried to_pickle() listed in dataframe documentation lacks property smaller corresponding .csv file holding same contents.

df = pd.get_dummies(pd.series(list('abca'))) df.to_csv("tmp.csv")       # 44 bytes df.to_pickle("tmp.pickle") # 682 bytes 

i'm looking fast (save time writing , reading in comparision .csv) , compressed (save disk space in comparision .csv) way (de)serialize pandas dataframes.

thanks.

pandas can export hdf5, faster , more compact csv. has problems mixed types, problems exist csv well.

you have careful comparing filesizes small datasets. overhead compression might make file larger smaller datasets