xlripper-test

Purpose

This repository serves as a submodule to xlripper which is a minimalistic xlsx parsing library in Go. The xlripper-test repo holds test files that are too large to ship with the main repo. The main repo can optionally use git submodule functionality to pull xlripper-test and expand the test suite to include these larger files.

Structure

Each xlsx test file must consist of these parts, at the root of the repository:

somefile.xlsx
somefile.meta.json
somefile.sheet0.csv
somefile.shhet1.csv

The test will automatically parse the xlsx file. It will use the information in somefile.meta.json to discover the sheet names. It will compare its extracted data and sheet names against the data provided in somefile.sheet0.csv (and additional sheets if they exist in the xlsx).

In the event that the test xlsx file is expected to fail parsing (for example, if the file is corrupt) then the json will indicate this and sheet.csv files are not necessary.

JSON Meta Structure

Each test xlsx file should include a json file with metadata. If the xlsx file is named myfile.xlsx then the metadata file should be named myfile.json.xlsx. The json file should look like this:

{
	"name": "myfile.xlsx",
	"isFailureExpected": false,
	"sheets: [
		"My Sheet 1",
		"Another Sheet Here"
	]
}

• name is the xlsx filename, string
• isFailureExpected should be true if the xlsx file is corrupted or unparseable, boolean.
• sheets A list of the sheets found in the xlsx file, array of strings in the correct order.

In the example above, we are declaring that the xlsx file has two sheets. Both of these sheets should be exported as separate csv documents using Microsoft Excel.

The test will then assert that myfile.sheet0.csv has the name My Sheet 1 and that we correctly parsed the data from that sheet.

Similarly the test will assert that myfile.sheet1.csv has the name Another Sheet Here and that we correctly parsed the data from that sheet.

Notice that the sheet filenames and index arrays are zero based even though programs like Microsoft Excel typically give then default names with 1-based suffixes.

Contributing

We need your help! We are looking for test failures of valid xlsx files from the wild. If you experience a failure with xlripper then this is the place to submit your failing file. Note that if you open your file with Microsoft Excel or another standard program in order to anonymize the data, then re-save it, you may strip the file of the uniqueness that caused it to fail.