I’m using pandoc (with the pandoc-crossref filter) to convert markdown documents to LATEX. But I want tables to be outputted as regular (floating) table environments rather than as (non-floating) longtables. I’ve written a LUA filter to do this, which is successful, except that it fails to extract the information about the table’s pandoc-crossref label. I need to extract this information so that I can include this label as part of the LATEX output (and thereby crossreference the table). Here’s a minimal example.

Firstly, here’s a minimal markdown document (input.md) containing just a table with the label tbl:table1 (which can be read by pandoc-crossref):

| First Header | Second Header |
|:-------------|:--------------|
| Content Cell | Content Cell  |
| Content Cell | Content Cell  |

: Table example {#tbl:table1}

Secondly, here’s my minimal LUA-filter (table.lua) that attempts to extract this label (tbl:table1) from this markdown document:

function Table(elem)
  local id = elem.attr.identifier or ''
  
  if id ~= '' then
    print("Table label: " .. id)
  else
    print("Table does not have a label")
  end
end

return {
  { Table = Table }
}

Thirdly, when I run pandoc from the command prompt, I expect this LUAfilter to print “Table label: tbl:table1” in the command prompt window, but instead it prints “Table does not have a label”. Here’s my command line in the command prompt:

pandoc input.md --output=output.tex --from markdown+pipe_tables --to latex --filter pandoc-crossref --lua-filter=table.lua

Any help greatly appreciated!

By the way, here is my wider LUA filter that (almost successfully) outputs a table rather than a longtable:

function Table(tbl)
  local simpleTable = pandoc.utils.to_simple_table(tbl)
  local blocks = pandoc.Blocks{}

  -- Function to create a LaTeX row
  local function create_latex_row(cells)
    local latex_row = ""
    for i, cell in ipairs(cells) do
      for _, content in ipairs(cell) do
        latex_row = latex_row .. pandoc.write(pandoc.Pandoc({content}), "latex")
      end
      if i < #cells then
        latex_row = latex_row .. " & "
      end
    end
    return latex_row .. " \\"
  end

  -- Determine the number of columns (using header or first row)
  local num_cols = 0
  if simpleTable.header and #simpleTable.header > 0 then
    num_cols = #simpleTable.header
  elseif #simpleTable.rows > 0 then
    num_cols = #simpleTable.rows[1]
  end

  -- Assume all columns are left-aligned for simplicity
  local col_alignment = string.rep("l", num_cols)

  -- Begin table environment
  blocks:insert(pandoc.RawBlock("latex", "\begin{table}[tbp]"))

  -- Add caption 
  if tbl.caption then
    local caption_text = pandoc.utils.stringify(tbl.caption)
    blocks:insert(pandoc.RawBlock("latex", "\caption{" .. caption_text .. "}"))
  end

  -- Add label using tbl.attr.identifier
  -- This is the part that doesn't work
  if tbl.attr and tbl.attr.identifier and tbl.attr.identifier ~= "" then
    blocks:insert(pandoc.RawBlock("latex", "\label{" .. tbl.attr.identifier .. "}"))
  end

  blocks:insert(pandoc.RawBlock("latex", "\centering"))
  blocks:insert(pandoc.RawBlock("latex", "\begin{tabular}{" .. col_alignment .. "}"))
  
  -- Process header
  if simpleTable.header and #simpleTable.header > 0 then
    local header_row = create_latex_row(simpleTable.header)
    blocks:insert(pandoc.RawBlock("latex", header_row))
  end

  -- Process rows
  for _, row in ipairs(simpleTable.rows) do
    local latex_row = create_latex_row(row)
    blocks:insert(pandoc.RawBlock("latex", latex_row))
  end

  -- End tabular and table environment
  blocks:insert(pandoc.RawBlock("latex", "\end{tabular}"))
  blocks:insert(pandoc.RawBlock("latex", "\end{table}"))

  return blocks
end