|
|
At 02:08 PM 3/9/2007, Ana Luebbers wrote:
>I'm using SPSS 14.0. for Windows and when I import a .csv file with
>200,000 records I always lose 250 to 300 records.
This is NOT the answer. I don't know the answer. (If it can't be
straightened out, maybe I'll ask you to E-mail your data, and I'll give
it a try.)
But there seem to be periodic reports of things like this. From
correspondence (I haven't observed or checked it out), at least
sometimes, it's not that the data aren't read; it's that the Data
Editor doesn't get fully updated immediately.
If you're judging from the data editor (or even SHOW N), try a
FREQUENCIES before being sure the cases are lost. Further, some of the
variables have lost the data while importing the file.
>Further, some of the variables have lost the data while importing the
>file.
THAT seems stranger, and I'm not aware of other relevant reports. It
may be helpful to know,
. Is it consistent, between successive attempts to load the same file?
That is, if you load the file and save it, then close SPSS, reopen it,
and load the file again, do you get exactly the same file? (Yes, I too
wish that SPSS made this easier to check.)
. Does it happen to numeric variables, string variables, or both?
. If numeric, does it happen to those with special formats, like dates?
. If string, do variables for which data are missing have the length
you'd expect?
. Does it seem to happen preferentially to variables near the
beginning, or near the end, of the records? In particular, do you ever
see variables read correctly, later in the record than variables whose
data is lost?
. When data is lost for a variable, is it all the data, or for only
some cases? (if the latter, OUCH!)
Good luck, and I'm afraid this is at best a beginning, not an answer,
Richard
|
Locked
