Handle _id columns in codelists for smoother use in sp_add_codelist()

[dan-bart]

dt<-sp_get_table("budget-central", 2017)
dt %>% sp_add_codelist("finmisto")
returns:
"Error: Something went wrong with matching the codelist to the data for period 2017-01-31.
Please inspect the dates on the codelist to make sure there are no duplicate items valid for one given date.
You may want to filter/edit the codelist manually and pass it to the add_codelist function as an object."

I had to rename the column and specify it in the function
finm<-sp_get_codelist("finmisto")
finm<-rename(finm, "finmisto" = "finmisto_id")
dt2<- dt %>% sp_add_codelist(finm, by = "finmisto")

Maybe it would be beneficial to change the colname in finmisto in "sp_get_codelist" function? (Unless it is used elsewhere)

[petrbouchal]

Thanks. I suspect that because the codelist contains these *_id columns, it is meant to be added in some other way.

I'd rather not rename codelists columns en masse to something not contained in the data until I understand why it is called X in data and X_id in the codelist. I think this might be a case of a secondary codelist as documented here
https://petrbouchal.xyz/statnipokladna/articles/statnipokladna.html#primary-and-secondary-codelists. I just don't know what this one is secondary to.

For this particular codelist, sp_add_codelist() actually joins on the ico column, which does not identify codelist rows uniquely, hence the error message.

For this particular codelist, you probably do want to join the _id and non-_id column, which can be done more concisely like so:

dt %>% sp_add_codelist("finmisto", by = c("finmisto" = "finmisto_id"))

(The by parameter is passed on to left_join() as is, so can be specified in the same way.)

But I found another genuine bug which might affect this, which is already fixed on Github - so just go ahead and use the Github version for now.

[petrbouchal]

(I updated the issue title to a more general one and will keep it open as a watching brief, in case I figure out the logic of the key column names)