This is an unsupported table hack. It adds mdtbl and mdtblio codeblocks. The mdtblio prints the code in a text area first before processing. mdtbls are easier to write than to read.

These table formats have been tested under pandoc 2.7 (by John MacFarlane).

The input file for this document is pandoctbl.md, which also includes the required CSS.

Setup

1. Download the pandoctbl perl code and this pandoctbl.md sample

2. Install the cpan module Perl6::Slurp.

3. Run pandoctbl pandoctbl.md -s -o pandoctbl.html (or pandoctbl pandoctbl.md -s --template ~/.pandoc/default.html4 -o pandoctbl.html).

CSS Used Here

<style>
  th { text-align:center; }
  .celll { text-align:left; }
  .cellc { text-align:center; }
  .cellr { text-align:right; }

  span.red { color:red; }
  span.blue { color:blue; }
  span.bgyellow { background-color:yellow; }
</style>

<style>
  th { text-align:center; }
  .celll { text-align:left; }
  .cellc { text-align:center; }
  .cellr { text-align:right; }

  span.red { color:red; }
  span.blue { color:blue; }
  span.bgyellow { background-color:yellow; }
</style>

Feature: Cell Alignment and Colspans

```{=mdtblio 1A} caption: table 1

—t ^ h1-left ^ h1-center ^ h1-right ^ h1-left-3col ^^^ h1-center-3col ^^^ h1-right-3col ^^^ —m | l1 | c1 | r3 | ctr4 | ctr5 | ctr6 | ctr7 | ctr8 | ctr9 | ctr10 | ctr11 | ctr12 | | l1 | c1 | r3 | ctr4-6 ||| ctr7-9 ||| ctr10-12 ||| | l1 | c1 | r3 | lft4-6-tres-cols ||| rght7-9-tres-cols ||| ctr10-12-tres-cols ||| —b

---

## Feature: Cell Content Preprocessing

```{=mdtblio 1A}
caption: table 2

---t
^                 ^                *Roots*                                      ^^^^
^ From X          ^    square    ^      cube       ^     tetra       ^      not    ^
---m
| For 2.0         |  1.414213562 |  1.259921050    |  1.189207115    |     -1.1487 |
| For 3.0         |  1.732050808 |  1.442249570    |  1.316074013    |     1.24573 | 
| For 4.0         |            2 |  1.587401052    |  1.414213562    |    -1.31607 |
| For 5.0         |     -2.23607 |  1.587401052    |  1.414213562    |         0.0 |
---m
| left            |        right |            both centered         ||       right |
^^                ^  bgyellow 2  ^     blue 1      ^       3         ^     cond 1  ^
---b

A format string, applied only to numbers in cells (strings are ignored), can be attached by starting a line with two carets (^^). A standalone number in the field is interpreted as digits after the decimal point. A string of at least two characters is interpreted as css. The “cond” string has special meaning: It attaches “blue” to positives, “red” to negatives, and “black to zero.

• column 2: bgyellow, two digits.
• column 3: blue, one digit.
• column 4: three digits.
• column 5: conditional (pos=blue,neg=red,zero=black), one digit.

Notes and Bugs

• This works by outputting both a =latex and a =html codeblock. The pandoc pass will process and then remove the useless one.

• The table format is very flexible.

  • You can abuse the left/center/right spacing to format individual cells in individual ways. Just redefine the latex macro and the html css. For example, you could define cellr not only to text-justify right, but also to italicize and overstrike and this would apply only in those cells that start with at least two spaces and end with exactly one space..

  • You could define a column style of ‘overstrike and have every number in such columns be overstruck via css.

• Rules in Tables

  • In html output, the table rules are simply replaced by emdashes according to the largest entry in all cells in this column. There is no intelligence about colspan spanning.

  • There is no support for clines. The obvious code would be ---2-4. The problem is deciding how many dashes should be added in html.

• LaTeX tables could be more intelligent in terms of nice output formatting.

• We want a realignment utility (for readability) md to md processor

• I am uncertain whether tabs should be expanded into spaces first.