Adding grouped sum of one variable to the original data - count

Question

My data look like:
Sample_ID Member_ID gender relative_ID relative_desc
1 11 male 1 Head
1 12 female 2 Partener
1 13 female 3 Child
1 14 female 3 Child
2 21 female 1 Head
2 22 male 3 Child
3 31 male 1 Head
3 32 female 2 Partener
4 41 male 1 Head
4 42 female 2 Partener
4 43 male 3 Child
4 44 male 3 Child
5 51 female 1 Head
5 52 female 3 Child
5 53 male 3 Child
5 54 male 3 Child
and many other columns..
what I want to know is how many child that each family has
and I did a lot of search and try to unpivot the relative_desc variable by :
COMPUTE Child = (relative_desc = "Child").
And then try to sum the aggregate with break of sample Id
DATASET DECLARE AggHouse.
AGGREGATE OUTFILE='AggHouse'
/BREAK SAMPLE_ID
/Child = SUM(Child).
this will move the sample id and number of child in each family into new dataset, what I did is merge the new sum column into the original dataset but I got a lot of missing data, any other suggestions?
thank you so much.

Answer 1

You can aggregate directly into the original dataset and save yourself work and troubles:
AGGREGATE OUTFILE=* mode=addvariables overwritevars=yes
/BREAK SAMPLE_ID
/Child = SUM(Child).
Note - the overwritevars subcommand lets you rewrite the Child variable with the sum. alternatively you could put the sum in a new variable like SumChild.
If you do prefer to aggregate to a new dataset and then reattach it back to the original dataset, please add to your post the syntax you used for that, so we can see what the problem was.