[Help] Mini.align - How to ignore delimiter if it's inside double quotes

[icyrainz]

As title, google sheet CSV format would put each column data as is if there is no comma inside and will wrap the column data inside double quote to differentiate with legitimate delimiter.

I have gone through the help file but could not a way to do this. Not sure if the plugin supports this capability or not.

Col1,Col2,Col2
Data1,Data2,Data3
"Data1WithComma,OtherSide",Data2,Data3

gAip,
-> The first column of the second entry should be "Data1WithComma,OtherSide" and not "Data1WithComma

[echasnovski]

Edit: Somewhat ignore this answer. There is a better built-in interactive way.

This question is relevant to me too, as I do quite frequently work with .csv files. But I don't think I found a way to organically fit this type of separator ignore into interactive mode of 'mini.align'.

There are two approaches here I can recommend:

• Manually filter out separators which you don't want to include. For your example it would be gAip, followed by f (built-in "filter" modifier) and typing not (row == 3 and col == 1) and <CR> (<Enter>). This will add filter pre-justify step which will exclude one separator from aligning.
• Write custom modifier which splits by commas not inside quotes. I had time and was curious and here is what I ended up with:

  local split_csv_row = function(row_string)
    -- Compute "forbidden" spans
    local forbidden = {}
    local add_to_forbidden = function(l, r) table.insert(forbidden, { from = l + 1, to = r }) end
  
    row_string:gsub("()%b''()", add_to_forbidden)
    row_string:gsub('()%b""()', add_to_forbidden)
  
    -- Compute all comma positions which are not inside forbidden spans
    local commas = {}
    row_string:gsub('%,()', function(r)
      for _, span in ipairs(forbidden) do
        if span.from <= r and r <= span.to then return end
      end
      table.insert(commas, r - 1)
    end)
  
    -- Split
    local n = #commas
    if n == 0 then return { row_string } end
  
    local res = { row_string:sub(1, commas[1]) }
    for i = 1, n - 1 do
      table.insert(res, row_string:sub(commas[i] + 1, commas[i + 1]))
    end
    table.insert(res, row_string:sub(commas[n] + 1))
  
    return res
  end
  
  local split_csv_step = require('mini.align').new_step(
    'csv-row',
    function(strings, _) return vim.tbl_map(split_csv_row, strings) end
  )
  
  require('mini.align').setup({
    modifiers = {
      c = function(steps, opts)
        steps.split = split_csv_step
        opts.split_pattern = '"csv-row"'
      end,
    },
  })

  Now you can gAip followed by c and it should split by commas ignoring spans inside '' and "".
  Please note, that this is not thoroughly tested and might need extra tweaks, but seems to work in the cases I tried.

[echasnovski]

Apparently, I forgot how my modules work 😞

There is an i modifier (for ignore) that does just that: adds matches inside "" to be ignored. So you can gAip followed by , to set separator and followed by i to add "ignore" pre-split step.

If you want to ignore inside both '' and "", you can customize i modifier like so:

local align = require('mini.align')
align.setup({
  modifiers = {
    i = function(steps, _) table.insert(steps.pre_split, align.gen_step.ignore_split({ '".-"', "'.-'" })) end,
  },
})

[icyrainz]

Thanks for the quick response. Works like a charm.