error in reshape long, "no xij variables found"

[Ann Miller]

I am trying to reshape a database from wide to long--I've done this hundreds of times successfully over the years but something is wrong in this particular dataset and I haven't managed to track it down. I am using Stata 15.1. I pared my data down to essentials to try to track the error and I still get the message that r(111); "no xij variables are found". I have 4 variables; record_id, sex, bt1 and bt2 which are scores on a test taken at 2 time points. I want to go to having record_id, sex bt and visit in one line instead of record, sex, bt1 and bt2. Note, I've also tried this after dropping "sex" from the database to just record_id and bt1 bt2 and get the same error message, so I think it must be something wrong in the ID or the bt variables. Any advice gratefully appreciated!


reshape long bt* sex, i(record_id) j(j)

* Example generated by -dataex-. To install: ssc install dataex
clear
input int record_id byte sex int bt1 float bt2
130 2 219 226
131 2 295 302
132 2 283 290
133 1 217 224
134 2 222 229
135 1 262 269
137 1 258 265
138 2 272 279
139 2 259 266
140 2 106 113
141 2  90  97
end
label values sex sex_
label def sex_ 1 "Masculino / Male", modify
label def sex_ 2 "Femenino / Female", modify

[Andrew Musau]

reshape long requires stubs of variable names, so you want

reshape long bt, i(record_id) j(j)

as you have bt1 and bt2, where the stub is "bt". See

help reshape

[Nick Cox]

reshape long bt, i(record_id) j(j)

reshape long wants to see stubnames, not variable names. Here I think btis the stub and sex just gets carried along.

[Ann Miller]

THANK YOU! I had forgotten that the "*" replaces the rest of the full name. Most grateful.

[Nick Cox]

Exemplary question that allowed a quick solution.