Add a DataFrame.show() method pls!

[halleygithub]

'print df' will give something like below if the dataframe 'df' is big to fit into the screen :

<class 'pandas.core.frame.DataFrame'>
MultiIndex: 41955 entries, (u'000002', u'20061231') to (u'603366', u'20120630')
Columns: 147 entries, STK_ID to EPS
dtypes: float64(135), object(12)

But most of the time, I want to have a glimpse of the data , which help to know what happened to the dataframe.

Can Pandas developers add a 'show()' method to DataFrame object to display part of the data inside ? Namely, show the four corner (up_left, up_right, down_left, down_right) data, and use '...' to represent the omitted part ?

somewhat like :

             STK_ID  RPT_Date STK_Name  ..  OprCF_PS    EPS

STK_ID RPT_Date
000002 20061231 000002 20061231 万科A .. -0.692 0.526
20070331 000002 20070331 万科A .. -0.741 0.140
20070630 000002 20070630 万科A .. -0.454 0.254
............... ............. ............... ........... ... .......... .....
20071231 000002 20071231 万科A .. -1.519 0.705
20080331 000002 20080331 万科A .. -0.207 0.105

[wesm]

Will keep it in mind-- happily accept a pull request, too, if you get around to it.

[halleygithub]

Below is the snippet that I currently use, pls be noted that not implement the row-wise function yet (the difficulty is that I don't know how to set/insert a row of '..'

def sw(df,first_rows = 20,last_rows =10,first_cols =3,last_cols =2):
''' display the df (can be dataframe or series) sample data
set_printoptions(max_columns=80,max_rows=30)
A,B (upt)
C,D (downpt)

'''
set_printoptions(max_columns=80,max_rows=30)

df =DataFrame(df) # convert to dataframe if input 'df' is series

ncol=len(df.columns)
nrow=len(df)

if ncol <= (first_cols + last_cols) :
    upt = df.ix[0:first_rows,:]         # screen width can contain all columns
    dowpt = df.ix[-last_rows:,:]
    pall = concat([upt,dowpt])
else:                                   # screen width can not contain all columns
    pa = df.ix[0:first_rows,0:first_cols]
    pb = df.ix[0:first_rows,-last_cols:]
    pc = df.ix[-last_rows:,0:first_cols]
    pd = df.ix[-last_rows:,-last_cols:]

    upt =  merge(pa,pb,how='inner',left_index=True, right_index=True)
    dowpt =  merge(pc,pd,how='inner',left_index=True, right_index=True)
    pall = concat([upt,dowpt])
    pall['..'] = '..'
    pall = __col_seq_set__(pall,['..'],[first_cols])

print "\n*****************************************************************"
print pall
print df.columns
print "row: %d    col: %d"%(len(df),len(df.columns))
print "*****************************************************************\n"
return None

DataFrame.show = sw
Series.show = sw

[changhiskhan]

@halleygithub you want to make this into a PR? You're almost there, just need to add a few test cases. Thanks in advance!

[halleygithub]

yes, pls feel free to further process as you want. I am a newbie to Pandas & github , not a programmer seriously. Feel good that I can contribute to the package.

[paulproteus]

It would be great for a contributor to take @halleygithub 's code here, add a test case, and submit it as a pull request.

[dundo4he]

Where is col_seq_set() ? Couldn't find it with grep -r "col_seq_set" pandas/*

[paulproteus]

@dundo4he it seems to me that it's a function that @halleygithub wrote and hasn't shared yet.

It's "probably" not too hard to figure out what it was, based on the output @halleygithub provided. Does that seem to be doable? If not, we should figure something else out.

[halleygithub]

def _col_seq_set(df, col_list, seq_list):
''' set dataframe col_list's sequence of 'df' by seq_list '''
df_col = list(df.columns)
fn_col = [x for x in df_col if x not in col_list]

for i in range(len(col_list)):
    fn_col.insert(seq_list[i], col_list[i])

return df[fn_col]

DataFrame.col_seq_set = _col_seq_set

[paulproteus]

Thanks, @halleygithub !

I'll just provide the same code with a preformatting tag:

def _col_seq_set(df, col_list, seq_list):                                                                                                                                                                  
    ''' set dataframe col_list's sequence of 'df' by seq_list '''                                                                                                                                          
    df_col = list(df.columns)                                                                                                                                                                              
    fn_col = [x for x in df_col if x not in col_list]                                                                                                                                                      
                                                                                                                                                                                                           
    for i in range(len(col_list)):                                                                                                                                                                         
        fn_col.insert(seq_list[i], col_list[i])                                                                                                                                                            
                                                                                                                                                                                                           
    return df[fn_col]                                                                                                                                                                                      
DataFrame.col_seq_set = _col_seq_set                                                                                                                                                                       

Also, @halleygithub , is it OK if we reuse your code under the same terms as pandas, available at https://github.com/pydata/pandas/blob/master/LICENSE ?

[halleygithub]

Sure, you can. I will feel good if I can help any. (Sorry for my ugly code :-) )

[halleygithub]

Oh, you just don't need "_col_seq_set()" at all, it is a function in my application to sort the columns sequence in batch. And in the "show()" method, you only need to put the pall[".."] column at first_cols+1 position intead of "pall = col_seq_set(pall,['..'],[first_cols])".

[dundo4he]

Please correct me if I am wrong. It seems that numpy.ndarray type automatically adjusts to fit the screen if the ndarray is too large. Can we borrow that mechanism?

import numpy as np

data = np.random.rand(100,100)

print data

[[ 0.98734521  0.54738576  0.43711897 ...,  0.11306541  0.22723003
   0.10952995]
[ 0.14806827  0.12672894  0.46958608 ...,  0.10808818  0.43853282
   0.02945122]
[ 0.8642931   0.40443047  0.93839959 ...,  0.70985694  0.99053461
   0.92551388]
 ..., 
[ 0.25710058  0.20474109  0.21222875 ...,  0.90249302  0.89936846
   0.14084486]
[ 0.04801022  0.85745347  0.76647051 ...,  0.85480267  0.23448934
   0.69833225]
[ 0.20308408  0.79021899  0.21764972 ...,  0.88353496  0.83787784
   0.82672697]]

[halleygithub]

yes, I also notice that . but I dislike numpy default format for

1. lot of '[' & ']'
2. no column name & index name (for Pandas dataframe)

[sinhrks]

Closed by #5550