glimpse-python

glimpse.py

def glimpse(df, maxvals=10, maxlen=110):
    print('Shape: ', df.shape)
    
    def pad(y):
        max_len = max([len(x) for x in y])
        return [x.ljust(max_len) for x in y]
    
    # Column Name
    toprnt = pad(df.columns.tolist())
    
    # Column Type
    toprnt = pad([toprnt[i] + ' ' + str(df.iloc[:,i].dtype) for i in range(df.shape[1])])
    
    # Num NAs
    num_nas = [df.iloc[:,i].isnull().sum() for i in range(df.shape[1])]
    num_nas_ratio = [int(round(x*100/df.shape[0])) for x in num_nas]
    num_nas_str = [str(x) + ' (' + str(y) + '%)' for x,y in zip(num_nas, num_nas_ratio)]
    max_len = max([len(x) for x in num_nas_str])
    num_nas_str = [x.rjust(max_len) for x in num_nas_str]
    toprnt = [x + ' ' + y + ' NAs' for x,y in zip(toprnt, num_nas_str)]
    
    # Separator
    toprnt = [x + ' : ' for x in toprnt]
    
    # Values
    toprnt = [toprnt[i] + ', '.join([str(y) for y in df.iloc[:min([maxvals,df.shape[0]]), i]]) for i in range(df.shape[1])]
    
    # Trim to maxlen
    toprnt = [x[:min(maxlen, len(x))] for x in toprnt]
    
    for x in toprnt:
        print(x)

This code tries to replicate the functionality of the 'glimpse' function from R package 'dplyr'

Beautiful! I was looking for something like that. Thank you!