Reputation: 11
I am working with a set of .dta files in Stata, each of which takes some time to create and each of which contains the date of creation in the file name (created at the point of saving using a macro with today's date).
At the moment my do-files identify the relevant .dta file to open based on the today's date macro, but this requires that I run the code to create the .dta files each day.
Is there a way of asking Stata to identify the most recently dated file from a set of files with same filename stem and different dates within a folder (and then open it), once I have run the "cd" command? I had a look on Statalist and SO but couldn't see an answer - any advice gratefully received.
e.g. In the folder, I have files 2020-08-23_datasetA.dta, 2020-08-22_datasetA.dta, 2020-08-22_datasetB.dta etc, and at different points I will want to select the most recently-dated version of A, B, C etc. Hence don't think a simple sort will work as there are datasets A, B, C at play.
(My question is essentially the Stata version of this one about R - Loading files with the most recent date in R)
[edited to clarify that there are multiple datasets, each of which is dated and each of which will need to be opened at different points]
Upvotes: 0
Views: 332
Reputation: 37208
Manifestly two or more files in a particular folder can't have the same name. But we know what you mean.
A utility like fs
from SSC will return a list of filenames matching a pattern, alphanumerically sorted. With your dating convention the last named will be the latest as your dates follow (year, month, day) order.
Using another convention for the rest of the filename won't undermine that, but naturally you need to spell out which subset of files is of interest. So a pattern is
. ssc install fs
. fs *datasetA.dta
. local wanted = word(r(files), -1)
where the installation need only take place once. You can circumvent fs
by using the calls to official Stata that it uses.
Perhaps you are seeking a program, but it's not clear to me that you need one.
Small detail: You're using the word macro in a way that doesn't match its Stata sense. Stata, for example, is not SAS! The terms code, routine and script are generic and don't conflict with any Stata use. Conversely, code, routine or script may have fixed meanings in other software you use. Either way, Stata questions are best phrased using Stata terms.
Upvotes: 1