asciitable, with column separation defined by divider rows

[velle]

It might be a rare case, but I have to parse a table that looks like this:

Name User group Email         Usertype
---- ---------- ------------- --------
cram cram       cram@fake.com null
jdoe jdoe       jdoe@fake.com null

Notice that the separation between columns is only 1 space, where most similar formatters use 2 spaces. This table format is the default format used by the NodeJS oclif table function, btw.

I realize that the current asciiparser parser cant solve this (that requires the column names to be separated by 2+ spaces). However, in case someone decides to make an even more generic parser, I wanted to leave this sample here, so that they can at least consider making it compatible for this format :)

Btw: I found jc today, and I think I love it :)

[kellyjonbrazil]

One idea is to pre-process the table before using the --asciitable parser in jc. This perl command nearly does the right thing by adding a space to every instance of one or more spaces in a line:

perl -pe 's/( +)/$1 . " "/ge'

It doesn't quite work with the example you gave because the group column is blank. If there is at least a character in the group column, then it seems to work. Anyway, just an idea:

% cat table.txt 
Name User group Email         Usertype
---- ---------- ------------- --------
cram cram -     cram@fake.com null
jdoe jdoe -     jdoe@fake.com null

% perl -pe 's/( +)/$1 . " "/ge' table.txt
Name  User  group  Email          Usertype
----  ----------  -------------  --------
cram  cram  -      cram@fake.com  null
jdoe  jdoe  -      jdoe@fake.com  null

% perl -pe 's/( +)/$1 . " "/ge' table.txt | jc --asciitable -p
[
  {
    "name": "cram",
    "user": "cram",
    "group": "-",
    "email": "cram@fake.com",
    "usertype": "null"
  },
  {
    "name": "jdoe",
    "user": "jdoe",
    "group": "-",
    "email": "jdoe@fake.com",
    "usertype": "null"
  }
]

[velle]

It doesn't quite work with the example you gave because the group column is blank

Its one column called "User group" :)

[kellyjonbrazil]

Ah, yes, this is why having only one space is ambiguous. It is difficult to tell whether it's a new column or just a multi-word column heading.

In this case, you can preprocess to turn "User group" into "User_group" and then do the perl trick to add a space between columns:

% cat table.txt | sed 's/User group/User_group/g' | perl -pe 's/( +)/$1 . " "/ge' | jc --asciitable -p
[
  {
    "name": "cram",
    "user_group": "cram",
    "email": "cram@fake.com",
    "usertype": "null"
  },
  {
    "name": "jdoe",
    "user_group": "jdoe",
    "email": "jdoe@fake.com",
    "usertype": "null"
  }
]