Add columns filter

The Add columns HXL filter expands a dataset with new columns, adding the same fixed value in every row.

For example, if you collect a 500-row dataset where every row applies to Kenya, you probably won't add a #country column and type "Kenya" in 500 times under it; however, when you exchange that dataset with others, they may need that information, since they may be working with datasets from several countries. Other fixed values — like the date or the reporting organisation — are also good candidates for this filter.

Options

The filter takes one or more new-columns specifications. The specification can be as simple as a regular HXL tag spec (with or without the leading #) followed by = and a value, like #org+reporting=CRS or date=2015-10-01. You can also include a new text header by placing it before the #; for example, this specification adds the tag #country and the text header "Country name": Country name#country=Kenya

TODO: row formulas

The filter also allows you to specify whether new columns should appear after existing columns (the default) or before them.

Example

Goal: add the date "2015-03-31" to the beginning of every row in a dataset. Use the new-column specification date=2015-03-31 and set the flag to place the column before existing ones.

Original dataset:

#org	#sector	#adm1
Org A	WASH	Coast District
Org B	Health	Mountain District
Org C	Education	Plains District
Org D	WASH	Coast District

Filtered dataset:

#date	#org	#sector	#adm1
2015-03-31	Org A	WASH	Coast District
2015-03-31	Org B	Health	Mountain District
2015-03-31	Org C	Education	Plains District
2015-03-31	Org D	WASH	Coast District

Usage

Command line

On the command line, use the hxladd program (hxladd -h for help):

hxladd -b -s 'date=2015-03-31'

Python

In a Python script, use the add_columns method:

hxl.data(url).add_columns('date=2015-03-31', before=True)

You can also include a list of new-column specifications:

hxl.data(url).add_columns(['date=2015-03-31', 'country=Kenya'], before=True)

JSON

To create an add columns filter in a JSON spec, use the following properties:

Property	Required?	Value
filter	yes	always "add_columns"
specs	yes	list of new-column specs (see above)
before	no	boolean

Example:

{
    "filter": "add_columns",
    "specs":  ["#country+name=Uganda", "#country+code=UGA"],
    "before": true
}