Alternative for selecting value in run_block calculation

[apriljunge]

Instead of

egress_block_A = filter(row -> any(occursin.(["1:main signal"], row.label)), run_block)[1,:s]

you cloud write

egress_block_A = run_block[run_block.label .== "1:main signal", :s][1]

For me the second option is more readable. What do you think?

[kaat0]

I am equally unhappy with both ways. 😅

My ideal in mind looks something like:
egress_block_A = get_poi(label = "1:main signal", run_block)[:s]

where TrainRuns.get_poi() returns a DataFrameRow.

However, I think we will need a better name for TrainRuns.get_poi().

[apriljunge]

Hm, I'm not quite sure if I would do this. You would have to deal with the flexibility of DataFrames. For example what if there are two rows with a label named 1:main signal.? If get_poi is not more than selecting one row by a label I would stick to the possibilities of the julia dataframe package. Extending Dataframes for this package would exceed the scope of this package.

As there are no real indices in Julia DataFrames one could use groupby like this

pois = groupby(run_block, :label)

egress_block_A = pois[(label="1:main signal",)][1,:s]
egress_block_B = pois[(label="2:main signal",)][1,:s]
...

[kaat0]

Good point!

For me, the groupby() function seems smart and sleek.