CSV enrichment¶
timeglyph csv reads a CSV and writes it back with timestamp columns rendered to
human-readable UTC. It works two ways, which can be combined.
Auto-detect (default)¶
With no other options, every numeric column is examined and, when it confidently looks like one timestamp format, a rendered column is appended to its right:
Auto-detection is deliberately conservative — it only enriches a column when:
- every non-empty cell is an integer at or above ~10⁸ (so counts, ids, and
years are skipped —
countabove is left untouched), and - the cells share one confident, in-window, non-sentinel top interpretation.
If a column is ambiguous, it is left alone rather than guessed — consistent with the engine's "never a single verdict" stance.
Explicit conversion¶
Convert a specific column under a chosen format with
--convert COLUMN:FORMAT (repeatable):
$ timeglyph csv files.csv --convert created:filetime --convert modified:unix
id,created,created_filetime,modified,modified_unix
1,132223104000000000,2020-01-01T00:00:00Z,1577836800,2020-01-01T00:00:00Z
The new column is named <column>_<format> and sits immediately to the right of
its source.
Replace in place¶
Use --replace to overwrite the source column instead of adding one:
Options¶
| Option | Effect |
|---|---|
| (none) | auto-detect numeric timestamp columns |
--convert COL:FMT |
convert column COL under format FMT (repeatable) |
--auto |
force auto-detect in addition to --convert |
--replace |
overwrite the source column instead of adding one |
-o, --output FILE |
write to FILE (default: stdout) |
path of - |
read the CSV from stdin |