This topic has been approached before in this forum but not, so far as I can tell, conclusively addressed.
I have ICD9 codes in string form. I need to export them in numeric form in order to upload them into a case-mix system. I am working with Stata for PC, v13.1. I have about 5 million diagnosis codes in my dataset. Here is what I have tried:
Destring: drops leading zeros
Encode: performs erratically. According to (http://www.stata.com/support/faqs/da...ring-variable/) I have too many unique values.
egen newdx = group(string_dx) Drops leading zeros
Changing the display will not work if you know how ICD9 codes are structured -- ICD9s are not all the same length and leading zeros are meaningful.
I have ICD9 codes in string form. I need to export them in numeric form in order to upload them into a case-mix system. I am working with Stata for PC, v13.1. I have about 5 million diagnosis codes in my dataset. Here is what I have tried:
Destring: drops leading zeros
Encode: performs erratically. According to (http://www.stata.com/support/faqs/da...ring-variable/) I have too many unique values.
egen newdx = group(string_dx) Drops leading zeros
Changing the display will not work if you know how ICD9 codes are structured -- ICD9s are not all the same length and leading zeros are meaningful.