I have a shapefile of census tracts and a CSV file with information I want to join to them. The CSV I am using has 65 rows, each for a unique census tract. I am using the census tract ID number as the key. For example of one the census tracts has an ID values of "1013.02". However, whenever I run the join I only get 60 of the rows from the CSV file to successfully join. I tried converting both key columns to double/real but that did not seem to help. I also checked to see if there were any duplicate values and there are none. I was curious if anyone has experience with a similar problem in the past? I have included links to a google drive folder containing the data and the website where I downloaded the census data from.



  • 3
    Have you tried converting the IDs to an integer? It's just a guess, but I suspect that will work better.
    – lambertj
    Dec 20, 2022 at 18:39
  • 1
    The problem with changing them to integers is that the numbers after the decimal place are critical parts of the ID number. For example I have IDs like 1115.41 and 1115.42.
    – Ian Moffit
    Dec 20, 2022 at 18:43
  • 3
    Would any IDs become non-unique if you just drop the period? So 1115.41 becomes 111541. Sorry, that's what I meant by integer, which was not very clear.
    – lambertj
    Dec 20, 2022 at 18:46
  • 4
    Multiply by 100
    – Babel
    Dec 20, 2022 at 19:02
  • 1
    So changing the IDs to integers did not help. When I went searched through the attributes of the census tract shape file I found 5 duplicate rows and 9 of the census tract values I was given in my CSV file were not present at all. Is there a place you all would recommend I download census tracts for Texas and the associated data from? Is there a way I can post the data that I have been using on Stack Exchange?
    – Ian Moffit
    Dec 21, 2022 at 1:56

1 Answer 1


So I was able to solve the problem. This time I tried downloading census tract shape file from this website:


I then used GeoID as the key field for the join rather than the tract number.

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service and acknowledge you have read our privacy policy.

Not the answer you're looking for? Browse other questions tagged or ask your own question.