columns are not properly aligned

[Rolf Leggewie]

Running datamash version 1.0.7-1 on Ubuntu trusty here. The output columns are not properly aligned (using "datamash -g 2 min 3 max 3 < /usr/share/doc/datamash/examples/scores.txt" for testing purposes). Please see attached screenshot.

[Assaf Gordon (agn1)]

Hello,

(I'm datamash's developer).

This is not a bug but intended operation: the 'scores.txt' is a tab-separated file.
datamash maintains the tabs as in the input - this is how it should be.

The output is not aligned on the terminal because some lines contain long text (longer than 8 characters) which then push the tab output to the next 'column'.

It is equivalent to the following situation:

  $ printf "a\t1\nvery-long-text\t2\n"
  a 1
  very-long-text 2

You can observe this with 'cat -A' (which displays tabs as '^I'):

  $ datamash -g 2 min 3 max 3 < scores.txt | cat -A
  Arts^I46^I88$
  Business^I79^I94$
  Health-Medicine^I72^I100$
  Social-Sciences^I27^I90$
  Life-Sciences^I14^I91$
  Engineering^I39^I99$

If you want to align tab-separated values properly on the screen, use 'expand':

  $ datamash -g 2 min 3 max 3 < scores.txt | expand -t20,30
  Arts 46 88
  Business 79 94
  Health-Medicine 72 100
  Social-Sciences 27 90
  Life-Sciences 14 91
  Engineering 39 99

regards,
 - assaf

[Erik Auerswald (auerswal)]

The 'column' utility from package 'bsdmainutils' can be used to properly align the columns:

[The web interface might "improve" the output by turning any run of space characters into a single space, removing the intended illustrative effect, as can be seen in the comment above]

    $ datamash -g 2 min 3 max 3 < /usr/share/doc/datamash/examples/scores.txt | column -s$'\t' -t
    Arts 46 88
    Business 79 94
    Health-Medicine 72 100
    Social-Sciences 27 90
    Life-Sciences 14 91
    Engineering 39 99

[With this specific input, one need not specify the separator, because no column comprises "blank" characters, i.e., the "-s" option can be omitted here, but not in general for the so called "TSV" format.]