v1.1.10: LDC 1.1.2; Statically linked Linux binaries.

NOTE: Pre-built binaries for this release are no longer available. Please use binaries from the latest release.

Changes in v1.1.10:

This release updates pre-built binaries to use the new LDC release, version 1.1.2. It also switches to static linking for the Linux pre-built binary. This makes the binaries usable on more Linux environments (issue #67). No functional changes.

v1.1.8 Version number printing

NOTE: Pre-built binaries for this release are no longer available. Please use binaries from the latest release.

Changes in v1.1.18:

No functional changes. All tools now print version information if given -V or --version options.

v1.1.7 Pre-compiled binaries for Linux and Mac

NOTE: Pre-built binaries for this release are no longer available. Please use binaries from the latest release.

Changes in v1.1.7:

No functional changes. Uses Travis-CI to produce pre-built binaries available from Github Releases. Binaries are generated for Linux and Mac using the latest LDC compiler.

Code coverage reports; minor cleanup

Code coverage reports are now available as part of automated tests.
A number of small cleanups in the code and documentation.

New tsv-summarize operator: quantile

The quantile operator enables calculating arbitrary percentiles when summarizing data. For example, the following command calculates the 10th, 25th, 50th, 75th, 90th percentile values for field 3 in file data.tsv:

$ tsv-summarize --quantile 3:0.1,0.25,0.5,0.75,0.9 data.tsv

Multiple fields can be specified. To calculate the 90th and 95th percentiles for fields 3, 4, and 5:

$ tsv-summarize --quantile 3,4,5:0.9,0.95 data.tsv

Quantile interpolation is done using the same method as the R default (interpolation method 7).

Other changes:

• Travis CI enabled.
• Additional detail added to the performance benchmark docs.

Bug fix: Add keep-header to dub builds

Bug fix: The new keep-header tool wasn't properly included in the dub build system. It was included in the make build system though. (Issue #44).

Documentation updates

Documentation changes related to the new keep-header tool.

New tool: keep-header

keep-header runs unix tools like sort and grep in a header-aware fashion. For example:

$ keep-header file1.txt -- sort

sorts the file, but the first line, presumably a header line, is output first and excluded from the sort. Multiple files can specified, only the header from the first file is included in the output.

Faster csv2tsv; Updated performance benchmarks

The csv2tsv tool was sped up 2x by using buffered writes rather than byte-at-a-time writes. The performance benchmarks were re-run to reflect the changes. csv2tsv is now the fastest tested tool on it's benchmark.

Publish updated performance benchmarks

No code changes, but a new set of benchmark results (see the docs/Performance.md).
Bumped the minor version number to reflect the changes in the last few months.