Just improvement for unit tests. We have some low coverage parts in around handling arrow record and parquet-go and these are so complicated.
before
$ make test
go test -race -cover ./...
ok github.com/reproio/columnify/avro (cached) coverage: 87.3% of statements
? github.com/reproio/columnify/cmd/columnify [no test files]
ok github.com/reproio/columnify/columnifier (cached) coverage: 8.0% of statements
ok github.com/reproio/columnify/parquet (cached) coverage: 65.9% of statements
ok github.com/reproio/columnify/record (cached) coverage: 61.8% of statements
ok github.com/reproio/columnify/schema (cached) coverage: 87.9% of statements
after
$ make test
go test -race -cover ./...
ok github.com/reproio/columnify/avro (cached) coverage: 87.3% of statements
? github.com/reproio/columnify/cmd/columnify [no test files]
ok github.com/reproio/columnify/columnifier 0.094s coverage: 88.0% of statements
ok github.com/reproio/columnify/parquet 0.087s coverage: 67.1% of statements
ok github.com/reproio/columnify/record 0.087s coverage: 65.1% of statements
ok github.com/reproio/columnify/schema 0.087s coverage: 92.4% of statements
Just improvement for unit tests. We have some low coverage parts in around handling arrow record and parquet-go and these are so complicated.
before
after