Option to only read n rows

[lexual]

If you have a large csv, can take quite some time to load up the file in tabview (10 seconds for file we just tested on)

Would be nice to have an option to only read in n rows.

e.g.

tabview -n 1000 my_big_data.csv

This would be useful as quite often I'm more interested in just looking at the structure and common values, not the entire dataset.

[firecat53]

That sounds reasonable. Thanks for the suggestion!

[s-celles]

It could be a good idea to support a sort of 'slice' syntax --rows 0:10 first 10 rows --rows -2: to get last 2 rows. same idea for columns with --cols. If data was in a Pandas DataFrame it will be very to have this slice concept.

import pandas as pd
import numpy as np
df = pd.DataFrame(np.random.rand(4,5), columns = list('abcde'), index=list('ABCD'))
print(df)

          a         b         c         d         e
A  0.393480  0.203721  0.502450  0.734149  0.380107
B  0.337915  0.416731  0.180090  0.840988  0.029033
C  0.094730  0.183898  0.875805  0.060895  0.387969
D  0.597270  0.697400  0.078505  0.850511  0.932793

You can get a subset of this dataframe (last 2 rows, first 3 colums)

print(df.iloc[-2:,0:3])
         a         b         c
C  0.09473  0.183898  0.875805
D  0.59727  0.697400  0.078505

[wavexx]

On 04/26/2015 10:23 AM, scls19fr wrote:

It could be a good idea to support a sort of 'slice' syntax --rows 0:10 first 10 rows --rows -2:0 to get last 2 rows. same idea for
columns with --cols. If data was in a Pandas DataFrame it will be
very to have this slice concept.

I'm a bit against this option, since you can now just chain a
head/tail/cut command in there, and it would actually be faster.

The only argument in favor is that it's a PITA to do sometimes, due to
the delimiters/quoting/whatnot.