Reshape with 3 variables that uniquely identifies the data.

[HeeSung Kim]

Dear Statalisters,

I have a dataset in long format as follows:

clear
input float(x y z data)
1 1 1  15
1 1 2 -10
1 1 3   7
1 2 1   9
1 2 2   3
1 2 3   0
1 3 1   8
1 3 2  13
1 3 3   6
2 1 1  21
end

Note that this is just a snippet of the example, and it continues in the pattern above. Thus my data is uniquely identified with 3 variables. I'd like to change this dataset into the following format:

clear
input float(x y data_1 data_2 data_3)
1 1 15 -10 7
1 2  9   3 0
1 3  8  13 6
2 1 21   . .
end

I'd have used reshape wide data, i(x) j(y) if it was only two variables, but now that there's 3 variables that uniquely identifies the data, I'm at a loss. I'd like to use by, but Stata doesn't allow by with reshape.

How can I change the dataset to wide? Any help will be appreciated.

[Rich Goldstein]

the following works for me (and thank you for using -dataex-)

reshape wide data, i(x y) j(z)

[HeeSung Kim]

That was perfect and simple! Never knew I could put more than one variable there. Thank you!