Column types for dot.

dotcols adds three things on top of dot: Num (running mean + stdev via Welford), Sym (counts table, mode, entropy), and Data (a CSV table that auto-builds one Num or Sym per column).

• Polymorphic. One add(c, x, train) call dispatches to num_add or sym_add via .c.is.
• One pass. Welford for Num, counts for Sym — both update in O(1) per row.
• Self-contained. The dotcols binary bundles all of dot's runtime plus these libs. One file install.
• Bundled demo. dotcols --demo stats data.csv prints per-column summary.

What it adds

# a single CSV ingest using dotcols
function header(    i) {
  for (i = 1; i <= NF; i++) {
    NAME[i] = $i
    COL[i]  = new($i ~ /^[A-Z]/ ? "num" : "sym") } }   # new from dot

function ingest(    i) {
  for (i = 1; i <= NF; i++) add(COL[i], $i, 1) }     # add from dotcols

END {
  for (i = 1; i <= length(NAME); i++)
    printf "%s\tn=%d\tmid=%s\tspread=%g\n",
      NAME[i], .COL[i].n, mid(COL[i]), var(COL[i]) }   # mid/var from dotcols

Same add(), mid(), var() calls work whether the column is Num or Sym. Dispatch is one line per op — concat type tag with op name, indirect-call via @fn.

Quick install + demo

curl -sL https://raw.githubusercontent.com/timm/awk/master/dotcols/dotcols -o dotcols
chmod +x dotcols
./dotcols --get-data       # 30 curated CSVs into ./data/
./dotcols --demo stats     # bundled stats demo on its sample.csv

Install · Stats walkthrough · Manual

Need ML on top? See dotlearn: trees, naive Bayes, active learning.