1
votes

Consider the following example dataset:

member like  deposit   age
1      1    9997    22
2      2     892    23
1      1     267    34
1      9    1728    54
3      1    9999    22
1      2    2123    34
9      2     445    13
1      1     928    22
1      1     276    34
2      1       .    23
2      1    3728    45
3      2    3652    34
.      1     451    35
.      1     231    67
2      9     234    17
3      2    3872    37
1      1     102    45
1      1     676    56
3      .       .    35
2      .    9999    67

All variables are numeric. The variable label and value labels are:

label var member "Are you a member of the club?"
label var like "Do you like it?"
label var deposit "How much in your account?"
label var age "Age"

label values member memberl, nofix
label define memberl 1 "member" 2 "nonmember" 3 "waiting" 9 "Refuse to answer"
label values like like, nofix
label define like 1 "like" 2 "don't like" 9 "don't know"
label value deposit dmoney, nofix
label define dmoney 9997 "N/A" 9999 "don't know"

The age variable has no system-missing values and no use-defined missing values. And there is a note (survey question number) for each variable:

notes member: QT35
notes like: QR22
notes deposit: Q6
notes age: info3

My goal is to export those information into a single data set (or a Excel table) as follows.

In a data set form:

http://i1279.photobucket.com/albums/y531/tpbest33/wanted_output_dataform_zps52953ecf.jpg

Or, in an Excel table form (this is preferable.):

http://i1279.photobucket.com/albums/y531/tpbest33/wanted_output_zps2c35208e.jpg

(Sorry, but I don't know how to write an html code to show the image.)

I am experimenting with a couple of Stata basic commands and packages: .uselabel, .labutil2, .valtovar., and others. Any help would be appreciated!

1
Not something I ever want to do, but I would not start from where you seem to be looking, which is at user-written extras. I would start with export excel. A wild guess is that you will have to copy-and-paste notes ad hoc. It is expecting too much to hope that Excel and Stata ideas will all line up one to one. - Nick Cox
Thanks, Nick. I managed to produce the table by using . descsave, .valtovar, . uselabel, and others. For notes, I made a dataset that has the variable and the corresponding note columns, and merge it onto the table I produced before. - Bill TP

1 Answers

1
votes

If you are comfortable working with JSON data, the jsonio program can export the entire dataset as a JSON object and/or write the JSON to a file. There is an option that also will print all of the meta data (e.g., variable labels, value labels, variable names, etc...) in the JSON object. If you wanted to see an example, you can view the README of the project's GitHub Repository. The README shows examples of what the output looks like using the auto.dta data set.