Difficulty reshaping data from wide to long format: variable id does not uniquely identify the observations

[Andrew Holm]

The purpose is to make the panel data ready for regression analysis, so I want to reshape it from wide to long format.

I tried the command:

reshape long pend, i(Kommunenummer) j(year)

I have no idea how to fix this. The column "Kommunenummer" are the individual observations while pend is a variable and the years are from 2000 to 2023. I'm trying to assess the impact of the pandemic on commuting.

The STATA file is attached below

https://drive.google.com/file/d/1oGS...usp=share_link

Any helpful inputs will be appreciated

[Nick Cox]

Did you type

 reshape error

as recommended?

Otherwise, type

duplicates list Kommunenummer

as implied (a little indirectly) by the FAQ https://www.stata.com/support/faqs/d...d-time-values/

One common problem is that you have a bundle of observations with missing values for commune number, as a side-effect of some spreadsheet nonsense.

Otherwise you may be being bitten by some accidental duplication or typos.

Your data aren't sorted by this variable. That's irrelevant to reshape but it makes a little difficult to see whether there are duplicates in the screenshot.

Note. I (and many others) are reluctant to download Stata datasets on grounds explained in the FAQ Advice.

Out of sight is a command that led to 2 missing values. That might be relevant.

[Andrew Holm]

Thank you @Nick Cox, the duplicates command worked, apparently the 2 last rows had missing values and were somehow counted, afterwards the reshape command transformed the data as desired. It seems the problem was deleting 2 rows from an Excel table and then importing that table into STATA.