Tabular: split taxonomy string

Tool

toolshed.g2.bx.psu.edu/repos/bgruening/text_processing/tp_awk_tool. This page covers the taxonomy-lineage awk recipe from iwc-tabular-operations-survey §2g and §7.

When to reach for it

Use this when a tabular column contains semicolon-delimited taxonomic lineage and downstream tools need explicit rank columns such as superkingdom, kingdom, phylum, class, order, family, genus, and species.

Do not use this as a generic pivot or transpose; see tabular-pivot-collection-to-wide for the collection-wide pivot pattern. Use tabular-compute-new-column for simple one-expression columns and tabular-cut-and-reorder-columns for post-split projection.

Parameters

• code: awk program. Corpus examples use multiline code for split, prefix dispatch, and missing-rank handling.
• infile: connected tabular input.
• variables: usually [] in awk examples.

Set FS = OFS = "\t" when reading and writing tabular rows.

Idiomatic shapes

Prefix-dispatch split, robust to missing or reordered ranks:

tool_id: toolshed.g2.bx.psu.edu/repos/bgruening/text_processing/tp_awk_tool/9.5+galaxy0
tool_state:
  code: |-
    BEGIN { FS = OFS = "\t" }
    NR > 2 {
      otu_id = "OTU_" $1
      superkingdom = kingdom = phylum = tax_class = order = family = genus = species = ""

      split($3, taxonomy, ";")

      for (i in taxonomy) {
        if (taxonomy[i] ~ /^sk__/) superkingdom = taxonomy[i]
        else if (taxonomy[i] ~ /^k__/) kingdom = taxonomy[i]
        else if (taxonomy[i] ~ /^p__/) phylum = taxonomy[i]
        else if (taxonomy[i] ~ /^c__/) tax_class = taxonomy[i]
        else if (taxonomy[i] ~ /^o__/) order = taxonomy[i]
        else if (taxonomy[i] ~ /^f__/) family = taxonomy[i]
        else if (taxonomy[i] ~ /^g__/) genus = taxonomy[i]
        else if (taxonomy[i] ~ /^s__/) species = taxonomy[i]
      }

      print otu_id, superkingdom, kingdom, phylum, tax_class, order, family, genus, species
    }
  infile: { __class__: ConnectedValue }
  variables: []

Anchored by the MAPseq-to-ampvis2 IWC exemplar.

Fixed-depth split when upstream guarantees positional ranks:

BEGIN { FS = OFS = "\t" }
NR > 1 {
  split($1, taxa_parts, ";")
  superkingdom = (taxa_parts[1] != "" ? taxa_parts[1] : "unassigned")
  kingdom = (taxa_parts[2] != "" ? taxa_parts[2] : "unassigned")
  phylum = (taxa_parts[3] != "" ? taxa_parts[3] : "unassigned")
  print superkingdom, kingdom, phylum
}

Pitfalls

• Set FS = OFS = "\t". Without it, print a, b, c emits spaces.
• Header skipping is source-specific. MAPseq examples use NR > 2; other table-reshaping examples use NR > 1.
• Choose prefix behavior deliberately. Decide whether downstream wants values like p__Firmicutes or stripped names like Firmicutes.
• Handle missing ranks explicitly. Corpus examples use empty strings or unassigned depending on downstream contract.
• for (i in array) order is undefined. It is fine for prefix dispatch, not for positional rank emission.
• Avoid class as a new example variable. Corpus-style awk may use it, but tax_class avoids confusing readers.

See also

• iwc-tabular-operations-survey — §2g, §2l, and §7 awk split decision.
• tabular-prepend-header — header manipulation before/after awk transforms.
• tabular-cut-and-reorder-columns — projection after rank expansion.
• tabular-pivot-collection-to-wide — collection-to-wide pivot, not taxonomy parsing.