SPSS incomplete file warning
Locked
 
|
I have a file which displays OK in the Data Editor, but I keep getting this message when I try to run an analysis.
Tried copying the file to drive e: but got same problem.
I can still get the raw data and the original SPSS files from Essex, but I’m intrigued as to why it won’t work. It was fine until today. Any ideas? John F Hall (Mr) [Retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com SPSS start page: www.surveyresearch.weebly.com/spss-without-tears.html |
Re: SPSS incomplete file warning
|
At 11:29 AM 5/24/2013, John F Hall wrote:
I have a file which displays OK in the Data Editor, but I keep getting this message when I try to run an analysis. That has the flavor of the file being physically corrupted -- that some (physical) records at the end got lost. It makes sense that the file opened OK in the data editor; I don't think the data editor reads more of the file than necessary to fill the display screen, so it wouldn't 'know' about what's missing at the end. Tried copying the file to drive e: but got same problem. That's what I'd have expected -- if records aren't in the (physical) file, they aren't. I can still get the raw data and the original SPSS files from Essex, but Im intrigued as to why it wont work. It was fine until today. Any ideas? It sounds like you've done that, and succeeded. Otherwise, look for backups of the file from before the last time you knew it worked, on the likelihood that some physical accident had corrupted the working copy, but an earlier copy might still be good. And here's code (untested) that might at least tell you how many records had been lost, and recover all those that are still there: GET FILE='C:\Users\John\Documents' + '\1 Social Research\Surveys' + '\British Social Attitudes' + '\bsa89\bsa89 SCPR version' + '\bsa89 essex version\' + 'gb8901.sav'. FILE HANDLE Recover /NAME='C:\Users\John\Documents' + '\1 Social Research\Surveys' + '\British Social Attitudes' + '\bsa89\bsa89 SCPR version' + '\bsa89 essex version\' + 'gb8901 RECOVER.sav'. SHOW N /* Display the number of cases that SPSS */ /* thinks should be in the file */. XSAVE OUTFILE=Recover. EXECUTE /* Required */. GET FILE=Recover. SHOW N. With good luck, that should create a valid file 'Recover' (see its full name in the FILE HANDLE statement) with all the cases that are present in the corrupted file. ===================== 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 |
|
|
Richard Many thanks for this: really helpful. I’ve managed to recreate the file, but I’ll try your solution to see what really happened. I take it you’ve seen my mails about the débâcle with SPSS 19 stopping suddenly in mid-flow, now thankfully resolved. John John F Hall (Mr) [Retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com SPSS start page: www.surveyresearch.weebly.com/spss-without-tears.html From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Richard Ristow At 11:29 AM 5/24/2013, John F Hall wrote: I have a file which displays OK in the Data Editor, but I keep getting this message when I try to run an analysis. Warnings Command: freq
Tried copying the file to drive e: but got same problem.
I can still get the raw data and the original SPSS files from Essex, but I’m intrigued as to why it won’t work. It was fine until today. Any ideas?
===================== 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 |
|
|
In reply to this post by Richard Ristow
Richard Just tried your syntax out in my first use of newly installed SPSS 21. Here’s what I got:
XSAVE OUTFILE=Recover. EXECUTE /* Required */. Error. Command name: EXECUTE Incomplete (truncated) SPSS Statistics data file: C:\Users\John\Documents\1 Social Research\Surveys\British Social Attitudes\bsa89\bsa89 SCPR version\bsa89 essex version\gb8901.sav (DATA1205) Execution of this command stops. Error # 1400. Command name: EXECUTE Input error when reading a case. Any changes made to the working file since 01-JUN-2013 05:57:11 have been lost. The time now is 05:57:13. GET FILE=Recover. Error # 63 in column 10. Text: Recover The file does not exist. Execution of this command stops. SHOW N. I made a new file on 29 April and ran SHOW on that: GET FILE='C:\Users\John\Documents\1 Social Research\Surveys\British Social Attitudes\bsa89jfh 29 April 05.sav'. DATASET NAME DataSet1 WINDOW=FRONT. SHOW N /* Display the number of cases that SPSS */ /* thinks should be in the file */.
Thanks for the suggestion, but I don’t think it’s worth any further effort to pin down the error. At least 21 works! John John F Hall (Mr) [Retired academic survey researcher] Email: [hidden email] Website: www.surveyresearch.weebly.com SPSS start page: www.surveyresearch.weebly.com/spss-without-tears.html From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Richard Ristow At 11:29 AM 5/24/2013, John F Hall wrote: I have a file which displays OK in the Data Editor, but I keep getting this message when I try to run an analysis. Warnings Command: freq
Tried copying the file to drive e: but got same problem.
I can still get the raw data and the original SPSS files from Essex, but I’m intrigued as to why it won’t work. It was fine until today. Any ideas?
===================== 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 |
||||||||||||||||||
Re: SPSS incomplete file warning
|
|
Dear John,
At 12:19 AM 6/1/2013, you wrote: Just tried your syntax out in my first use of newly installed SPSS 21. Not at all what I'd expected. I wonder how SPSS does look up 'N'? I don't think it reads the entire file to do it, so I'd assumed it's in the dictionary data somewhere. Might be worth a post asking about this, noting that it's obtained after loading a .SAV file known to be damaged. XSAVE OUTFILE=Recover. This, I'd more or less expected: the transformation program (containing only the command XSAVE) terminates on an error when it fails to find all the cases it expects. (Which means it must somewhere have recorded how many to expect.) So far, as expected. But you also got, GET FILE=Recover. Bother. I'd hoped that SPSS would build the output file written to by XSAVE (indeed, surely it does), and then close it properly when the error arose. Doesn't look like it does. Pity. Might be worth commenting on this in a posting, too. T hanks for the suggestion, but I dont think its worth any further effort to pin down the error. No, unless it's worth a posting inquiring about how SPSS does handle this abnormal situation -- first, how it can load a file at all and not have SHOW N work; second, what action is taken when the error arises while writing a file using XSAVE. But that's for general interest. I can't claim it would forward your work. I'm glad your work is going OK. -Best regards, Richard ===================== 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 |
| Free forum by Nabble | Edit this page |