btskinner / btskinner.github.io Goto Github PK

Hi Ben,

I know we talked about this earlier year, and you said I should submit a bug report for it. I couldn't find a repo for merging ipeds, so I figured I'd submit it here.

With some of the obscure finance datasets that we were using from IPEDS, the file name has 5 digits in it (e.g. F1718_F1A), so when merging ipeds runs the year comes out messed up. So e.g. for F1718_F1A the year ends up coming out as 2172

Original (Lines 79-83)

get year from file name

    year <- as.numeric(gsub("\\D", "", zfiles[i]))
    ## convert split year (e.g., 0910 to 2009)
    if (year < 2000) {
        year <- round(year/100, digits=0) + 2000
    }

The potential fix I came up with is to, instead of deleting everything that is not a digit using gsub, extract the first 4 digits from the file name with str_extract

Potential Fix (Lines 79-83)

get year from file name

year <- as.numeric(str_extract(zfiles[i], "\\d{4}"))
## convert split year (e.g., 0910 to 2009)
if (year < 2000) {
  year <- round(year/100, digits=0) + 2000
}

Hopefully that's useful! I've changed my project script to use this fix (as opposed to the case when we had been using) and it seems to work 🤞

Matt