seqfu tabcheck

Validate TSV/CSV files by checking consistent field counts across rows (gzipped input supported).

seqfu tabcheck is the preferred command.
The legacy binary fu-tabcheck is still available and accepts the same options.

tabcheck

A program inspect TSV and CSV files, that must contain more than 1 column.
Double quotes are considered field delimiters, if present.
Gzipped files are supported natively.

Usage:
  tabcheck [options] <FILE>...

Options:
  -s, --separator CHAR   Character separating the values, 'tab' for tab and 'auto'
                         to try tab or commas [default: auto]
  -c, --comment CHAR     Comment/Header char [default: #]
  -i, --inspect          Inspect one valid table and infer column types and statistics
  --header               Print a header to the report
  --verbose              Enable verbose mode

Output

Regular mode prints one line per file with pass/fail, detected separator, and row/column counts.

On failure, diagnostics include:

  • first bad row index
  • expected number of columns
  • observed number of columns

Example error:

data/table2.tsv    Error[row=3;expected=3;observed=4;reason=inconsistent-column-count]

Notes

  • --separator auto performs separator sampling before full parsing.
  • Comments are ignored when --comment is set (default: #).
  • --inspect accepts exactly one table and validates it before printing a profile.
  • A leading comment/header row supplies column names. An unmarked first row is inferred as a header when its labels differ from numeric or date values below; otherwise columns are numbered from 1.
  • Types are inferred from all non-empty values as int, float, date, or string. Dates recognize YYYY-MM-DD and YYYY/MM/DD.
  • Numeric descriptions report minimum, maximum, and average across non-empty values. String descriptions report total and distinct counts plus the three most frequent values and their percentages. Date descriptions report the minimum and maximum date.

Examples

Validate two files:

seqfu tabcheck data/table.tsv data/table.csv

Inspect column profiles:

seqfu tabcheck --inspect --header data/table.tsv

The profile is tab-separated:

Column  Type    Description
sample  string  total=10; distinct=3; top3=A: 5 (50.0%), B: 3 (30.0%), C: 2 (20.0%)
count   int     min=1; max=12; average=6.4

Legacy equivalent:

fu-tabcheck data/table.tsv