Columns

A series step brings many columns. These three steps shape them.

select

Keeps only the columns given, in the order given.

# A list keeps the names.
select: [year, marz, value]

# A mapping renames as it keeps: new: old.
select:
  year: year
  wage: value

rename

Renames columns and keeps every column where it is: old: new.

rename: { value: wage, series: indicator }

drop

Removes the columns given and keeps the rest in order.

drop: [series_id, lat, lon]

Example: pull out a series and name its columns

nodes:
  wages:
    series: { find: "average monthly nominal wages", place: Yerevan }
    resample: { to: year, how: mean }
    select: { year: year, wage_amd: value }

The result has exactly two columns, year and wage_amd.

Two columns cannot end up with the same name; a column that does not exist is an error that lists the columns there are.