At 11:05 AM 1/8/2009, Gene Maguin wrote:
>I get an entry consisting of three dots interspersed with valid
>values in a frequencies command listing. Actual example:
>
>Value Frequency, etc.
>-0.009075342465753422 1
>... 1
>... 1
>0.333750521485196 1
>
>Causes?
Well :-[ , I probably could give a pretty good argument that it didn't happen.
("Ah! What avails the classic bent, And what the cultured word,
Against the undoctored incident That actually occurred?")
The first remark is that you normally don't run Frequencies on data
like this, with values that don't look very discrete. But if you're
doing it, you've got some reason.
Wild guess: Could the formatted form of the values that aren't
displayed, be longer than some limit in Frequencies for the length of
the value?
What happens if you do a crude synthetic Frequencies, like this (untested)?
AGGREGATE OUTFILE = *
/BREAK = <whatever variable you're running Frequencies on>
/COUNT = NU.
LIST.
You could also try changing the format to PIBHEX16. I *believe* that
gives the exact internal representation of the number -- admittedly,
difficult to hand-decode afterward.
=====================
To manage your subscription to SPSSX-L, send a message to
[hidden email] (not to SPSSX-L), with no body text except the
command. To leave the list, send the command
SIGNOFF SPSSX-L
For a list of commands to manage subscriptions, send the command
INFO REFCARD
Locked
