recode to exclude missings without specifying an upper limit?

[Alison El Ayadi]

Dear Statalisters -

I am looking to recode a numeric variable where the highest category does not specify an upper value but instead indicates that the category should include all values above X but exclude missings (or less than missing). Is there an easy way to do this within the recode statement?

Thank you,
Alison

[Nick Cox]

Hmm. Suppose 42 means "42 and above". What would you recode it to?

[Alison El Ayadi]

For example, my code is as follows for recoding age into age categories:
recode age (0/17=1 "1:Age 0-17") (18/44=2 "2:Age 18-44" ) (45/100=3 "3:Age 45+"), gen(age_gr) label(age_gr)
Where I would prefer not to specify an upper age limit for the final category 45+ as I have been surprised by a few exceptions (101, 102 . . . ).
There is of course a component of biological plausibility within this example, but it has come up for me in the past with other variables.

[Svend Juul]

You can use min and max, max meaning the highest nonmissing value:

recode age (min/17=1 "1:Age 0-17") (18/44=2 "2:Age 18-44" ) (45/max=3 "3:Age 45+"), gen(age_gr) label(age_gr)

[Alison El Ayadi]

Thank you, Svend! I was not aware of min/max usage.