miller/test
John Kerl 1a20b32bb5
join: add --ignore-empty to skip pairing on empty-string join keys (#1194) (#2210)
By default, records with an empty-string value in a join field are
paired just like any other value, so two records both missing an ID
get matched with each other -- rarely the intended behavior.
--ignore-empty treats an empty-string join-field value as if the
field were absent, on both the left and right files, so such records
fall through to unpaired handling (--np/--ul/--ur) instead of
cross-joining on the empty string.

Wires the check through both the default half-streaming join and the
-s/--sorted-input doubly-streaming join, which track left-file
buckets independently and needed the same empty-aware key-presence
check in JoinBucketKeeper.

Co-authored-by: Claude Sonnet 5 <noreply@anthropic.com>
2026-07-15 16:42:39 -04:00
..
cases join: add --ignore-empty to skip pairing on empty-string join keys (#1194) (#2210) 2026-07-15 16:42:39 -04:00
cases-not-suitable-for-ci
expected
input join: add --ignore-empty to skip pairing on empty-string join keys (#1194) (#2210) 2026-07-15 16:42:39 -04:00
stdlib
README.md

Miller regression tests

There are a few files unit-tested with Go's testing package -- a few dozen cases total.

The vast majority of Miller tests, though -- thousands of cases -- are tested by running scripted invocations of mlr with various flags and inputs, comparing against expected output, and checking the exit code back to the shell.

How to run the regression tests, in brief

Note: while this README.md file is within the test/ subdirectory, all paths in this file are written from the perspective of the user being cd'ed into the repository base directory, i.e. this directory's parent directory.

  • mlr regtest --help

  • go test github.com/johnkerl/miller/v6/pkg/... — runs the Go unit tests (a few dozen cases).

Items for the duration of the Go port

  • mlr regtest -c ... runs the C version of Miller from the local checkout

More details

You can alias mr='mlr regtest' for convenience. With no arguments, mr runs all cases under test/cases/. Pass one or more paths to run only those directories or specific .cmd files.

  • mr — run all regression cases (default path is test/cases/).
  • mr test/cases/foo — run only cases under that directory.
  • mr -v test/cases/foo — same, plus per-command pass/fail; use -vv or -vvv for more detail.
  • mr -j test/cases/foo/0003 — show the Miller command, any script, and actual output for that case (handy for debugging).
  • mr -p test/cases/foo/0003populate: write or overwrite expout and experr from the current run (use when adding or updating expected output).
  • mr -c ... — use the C build of Miller (e.g. -c../c/mlr) instead of the current executable.

To review populated files before committing, run mr -p on the desired path, then git diff to inspect changes and git reset --hard to discard them.

Creating new cases

  1. Create a case directory under test/cases/, e.g. test/cases/my-feature/0001.
  2. Add a cmd file containing the Miller command line (one line), e.g. mlr cat test/input/simple.dkvp.
  3. Use shared input under test/input/, or add a local input file in the case directory; in cmd you can use ${CASEDIR} so the command refers to the case directory (e.g. mlr cat ${CASEDIR}/input).
  4. Run mlr regtest -p test/cases/my-feature/0001 to generate expout (and experr if the command produces stderr). If the command is expected to exit non-zero, add an empty should-fail file.
  5. Run mlr regtest test/cases/my-feature/0001 (without -p) to confirm the case passes.

Optional: mlr — DSL script file when the test uses -f/put/filter; env — environment variables to set for the case (unset after).