[SOLVED] How to subset rows with an OR condition in Julia DataFrames

How to subset rows with an OR condition in Julia DataFrames

I have a DataFrame and I want to filter the rows where column During_Cabg OR column During_Pci have a value of 1. Here's what I'm doing:

pci_or_cabg = @chain df begin
    select([:During_Cabg, :During_Pci] .=> ByRow(x -> coalesce.(x, 0)); renamecols=false)
    subset(:During_Cabg => ByRow(==(1)), :During_Pci => ByRow(==(1)))
end

The problem is that this line: ByRow(==(1)), :During_Pci => ByRow(==(1) seems to imply an AND not OR. The result I'm getting is values where BOTH columns are 1 (not what I want).

How to subset a DataFrame with multiple conditions (AND or OR) with multiple columns?

Thank you!

Solution

In subset AND condition is used. If you want to use OR you need to pass it in a single condition. A general way to do it would be:

subset(df, AsTable(columns) => ByRow(x -> any(predicate, x)))

subset(df, columns => ByRow((x...) -> any(predicate, x)))

if you want to apply the same predicate to all columns.

If you want a shorter syntax consider using one of the meta-packages, for example with DataFramesMeta.jl you can write):

@rsubset(df, :During_Cabg == 1 || :During_Pci == 1)

which works nice with chaining.