Data management on flat data files

[Eugenia Cho]

Hi all-

I need to merge, clean and label covariates on few flat data files from the CDC NCHD Vital Statistics dataset, specifically the 2022 period/2021 cohort. I converted them to txt format and manage to import them into Stata 18. However, the imported data appeared as one column with 3,669,929 observations. I need to label the covariates based on the data dictionary but need some advice on how to do this. Appreciate your advice, thanks!

import delimited "/Users/yc765/Downloads/VS2021LINK.Public.USDENPUB_add_Seq_for2021Cohort_r 2024_04_15.txt",	emptylines(include)
> clear
(encoding automatically selected: UTF-8)
(1 var, 3,669,929 obs)

[Eugenia Cho]

I try to use dataex to cut out a sample of the data as example to show here, but because the width of the column is too wide, I get an error message that the line size was exceeded...

[Bert Lloyd]

It looks like you will have to use infile with a dictionary based on the "File Layout" provided starting on page 22 of the PDF you link.

If a Stata dictionary file for that particular dataset does not exist you may have to code it yourself.

See

https://www.stata.com/manuals/dinfil...le(fixedformat)

It looks like there are some other useful guides and examples online such as

https://www.stata.com/support/faqs/d...d-format-data/

and

https://www.stata.com/support/faqs/d...onary-options/


It looks as though NBER created dictionary (.dct) files through 2011, those may be helpful if it turns out you need to write your own dictionary file:

https://www.nber.org/research/data/l...th-cohort-data

[Eugenia Cho]

Thank you Bert, I'll try your suggestion.

[Erik Reinbergs]

Have you looked at whether this resource would work? https://www.nber.org/research/data/m...use-death-data
Unfortunately it looks like they don't have the 2022 data yet?

[Eugenia Cho]

Erik, you are right. No 2022 data yet on NBER website, thanks though.

[Eugenia Cho]

Bert Lloyd Just an update- I've to create the dictionary using the infile command. The .dta file though needs to be changed to .raw in order for .dct to work with infile. Hope this will be helpful for someone else doing the same in the future. Thanks again.