v1.0.3

Fix value truncation for expanded floats whose string representation exceeded 256 digits.

v1.0.2

Handle number values with leading + characters.

v1.0.1

Catch invalid date values that are too short before passing to timestamp validation to prevent index out-of-bounds

v1.0.0

Initial "Complete" Release

Changes

• Input data is now expected to be in a headerless 3 column TSV format with the columns "attribute_stable_id", "entity_id", "value".
• Input may now be a list of files rather than just one
• Added an option to indicate that inputs are already sorted/grouped by the attribute_stable_id and entity_id columns. When this flag is present or set to true the tool will process the data as a stream, when the flag is omitted or set to false the input will be collected and sorted in memory before data processing will begin.

v0.5.3

Fixes

Upper and lower quartile calculations now handle small sets of input values.

v0.5.1

• Update dependency.

v0.5.0

• Update script dependencies.

v0.4.0

Changes

• Add cli flag --format | -f which may be one of tsv|csv|json. Defaults to tsv.
• Add cli flag --headers | -t which causes the tool to print out a header line. Only applies to tsv and csv output formats. Defaults to false.

v0.3.0

Now returns the following stats as a tab-delimited string in the order defined here:

1. min
2. max
3. binWidth
4. mean
5. median
6. lower_quartile
7. upper_quartile

v0.2.0

Adds:

• support for boolean values matching the initial R implementation which simply returns an NA value
• the rm.na variable as an optional cli flag which defaults to false
• optional file path argument to read from file instead of stdin

WARNING The default behavior is changed due to the addition of the --rm-na(-r) flag. NA values are now not removed from the input by default, meaning if the input contains one or more NA values (empty strings) the output will be an NA value (empty string).