We are currently running SPSS for windows - versions 6, 13 and 14.
We would prefer to run version 13 or 14, rather than version 6. However, we prefer the version 6 output format for tables. Does anyone have any experience of converting version 13/14 output (tables only) to version 6 format? I have tried exporting in draft format and then post-processing (using Perl) to convert to version 6 table format, but without total success. Is there an easier route? Any help would be much appreciated! Elaine -- ***************************************** Elaine Prentice-Lane ([hidden email]) ISER, University of Essex, Wivenhoe Park, Colchester, Essex CO4 3SQ Tel: +44 (0)1206 873402 Fax: +44 (0)1206 873151 ***************************************** |
The Draft Viewer, as you probably know, comes closest to the old output format, but it will not reproduce the actual structure of v6 tables, since the output was reorganized after that.
What is it exactly that you need about v6 format? If you export the regular Viewer output as XML or use OMS to save XML, you can apply XML tools such as XSLT to transform the output into anything you want, including reorganizing it, much more effectively than applying Perl to plain text. And you could then have the output as plain text, html, or other formats. There are many free XSLT engines available, but you do have to learn a different programming model to use it effectively. HTH, Jon Peck -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Elaine Prentice-Lane Sent: Wednesday, January 17, 2007 7:04 AM To: [hidden email] Subject: [SPSSX-L] SPSS output - Versions 6 -> 13 or 14 We are currently running SPSS for windows - versions 6, 13 and 14. We would prefer to run version 13 or 14, rather than version 6. However, we prefer the version 6 output format for tables. Does anyone have any experience of converting version 13/14 output (tables only) to version 6 format? I have tried exporting in draft format and then post-processing (using Perl) to convert to version 6 table format, but without total success. Is there an easier route? Any help would be much appreciated! Elaine -- ***************************************** Elaine Prentice-Lane ([hidden email]) ISER, University of Essex, Wivenhoe Park, Colchester, Essex CO4 3SQ Tel: +44 (0)1206 873402 Fax: +44 (0)1206 873151 ***************************************** |
In reply to this post by Elaine Prentice-Lane
Why do you prefer tabular output from version 6? There are a number of ways of converting output so what you are after will dictate your solution.
-----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Elaine Prentice-Lane Sent: Wednesday, January 17, 2007 7:04 AM To: [hidden email] Subject: SPSS output - Versions 6 -> 13 or 14 We are currently running SPSS for windows - versions 6, 13 and 14. We would prefer to run version 13 or 14, rather than version 6. However, we prefer the version 6 output format for tables. Does anyone have any experience of converting version 13/14 output (tables only) to version 6 format? I have tried exporting in draft format and then post-processing (using Perl) to convert to version 6 table format, but without total success. Is there an easier route? Any help would be much appreciated! Elaine -- ***************************************** Elaine Prentice-Lane ([hidden email]) ISER, University of Essex, Wivenhoe Park, Colchester, Essex CO4 3SQ Tel: +44 (0)1206 873402 Fax: +44 (0)1206 873151 ***************************************** |
In reply to this post by Elaine Prentice-Lane
At 08:04 AM 1/17/2007, Elaine Prentice-Lane wrote:
>We are currently running SPSS for windows - >versions 6, 13 and 14. We would prefer to run >version 13 or 14, rather than version 6. >However, we prefer the version 6 output format for tables. Like Jon Peck and ViAnn Beadle, I'd be interested in what you like better about version 6 output. In my case, because I feel somewhat the same - in ten years I haven't completely recovered from the shock when output changed from listing form (v.6 and before), and pivot tables (v.7 and after). <<To my friends at SPSS, Inc.: apologies for not having found a comfortable way to join the bandwagon.>> When you say 'tables', what do you mean? That is, generated by what commands or procedures? Most of what I've had trouble with is in presenting output outside SPSS. To my mind, the viewers, especially the output viewer, *are* viewers, meant to be seen, only very second to be exported. I've struggled. Sometimes exporting to HTML has done all right, though I've had to do (crude) editing to shorten long output lines to reasonable. Sometimes draft output works best. It's far the best way I've found to put output in E-mail, certainly, especially text-only E-mail -- you can see program listings in my postings prefixed "SPSS draft output." (And at that, it's edited.) Exporting: I liked to print output, punch three holes in it, and put it in ring binders. Listing-format SPSS output files were pretty good for this: with page breaks in logical places, and TITLE, SUBTITLE, and FILE LABEL in the page headers. Neither output (.spo) nor draft output (.rtf) are very good for that. For program code, I print the .rtf, but it isn't very good for it: there are SPSS page breaks, but they aren't output as .rtf page breaks; of course (is it of course?), titles and the like aren't output as .rtf page heads. Anyhow, I'm a pack rat. I don't run SPSS 6 now, but I've kept listings from at least one project. By way of contrast, here's a snippet from one of the listings (cheating a little-changing the box characters to ones that look good with standard font. The '+' in the first position marks where SPSS put in a page break: -> TITLE -> 'LLC 6/23/97 Q.3: Age distributions -- sample, age 0-17'. -> SUBTITLE -> '(WRR, interactive, 6/25/97)'. -> CROSSTABS -> /TABLES=ageyear BY sex -> /FORMAT= AVALUE NOINDEX BOX LABELS TABLES -> /CELLS= COUNT ROW COLUMN TOTAL . Memory allows for 11,915 cells with 2 dimensions for general CROSSTABS. +25 Jun 97 LLC 6/23/97 Q.3: Age distributions -- sample, age 0-17 Page 10 16:04:44 (WRR, interactive, 6/25/97) File: CENS_MRG: Sample & population individual demographics (WRR) AGEYEAR Age, to lower whole year (FOSTAUD3) by SEX Sex of individual (FOSTAUD3) SEX Page 1 of 2 Count | Row Pct |Male Female Col Pct | Row Tot Pct | 1 | 2 | Total AGEYEAR --------+--------+--------+ 0 | 10 | 11 | 21 | 47.6 | 52.4 | 5.5 | 5.1 | 6.0 | | 2.6 | 2.9 | +--------+--------+ 1 | 13 | 13 | 26 | 50.0 | 50.0 | 6.8 | 6.6 | 7.1 | | 3.4 | 3.4 | +--------+--------+ 2 | 17 | 14 | 31 | 54.8 | 45.2 | 8.1 | 8.6 | 7.6 | | 4.5 | 3.7 | +--------+--------+ 3 | 17 | 23 | 40 | 42.5 | 57.5 | 10.5 | 8.6 | 12.5 | | 4.5 | 6.0 | +--------+--------+ 4 | 15 | 19 | 34 | 44.1 | 55.9 | 8.9 | 7.6 | 10.3 | | 3.9 | 5.0 | +--------+--------+ 5 | 19 | 11 | 30 | 63.3 | 36.7 | 7.9 | 9.6 | 6.0 | | 5.0 | 2.9 | +--------+--------+ 6 | 15 | 10 | 25 | 60.0 | 40.0 | 6.6 | 7.6 | 5.4 | | 3.9 | 2.6 | +--------+--------+ 7 | 14 | 8 | 22 | 63.6 | 36.4 | 5.8 | 7.1 | 4.3 | | 3.7 | 2.1 | +--------+--------+ 8 | 11 | 12 | 23 | 47.8 | 52.2 | 6.0 | 5.6 | 6.5 | | 2.9 | 3.1 | +--------+--------+ 9 | 7 | 10 | 17 | 41.2 | 58.8 | 4.5 | 3.6 | 5.4 | | 1.8 | 2.6 | ---------------- Column 197 184 381 (Continued) Total 51.7 48.3 100.0 ............................................... And, since I've also kept the code and the saved file, here's the same as SPSS 14 draft output. I grant there are reasons for all the changes, and that it could be shortened quite a lot by omitting Notes: TITLE 'LLC 6/23/97 Q.3: Age distributions -- sample, age 0-17'. SUBTITLE '(WRR, interactive, 6/25/97)'. CROSSTABS /TABLES=ageyear BY sex /FORMAT= AVALUE NOINDEX BOX LABELS TABLES /CELLS= COUNT ROW COLUMN TOTAL . Crosstabs Notes |-----------------------------|---------------| |Output Created |17-JAN-2007 | | |17:06:50 | |-----------------------------|---------------| |Comments | | |-------------|---------------|---------------| |Input |Data |c:\documents | | | |and | | | |settings\richar| | | |d\my | | | |documents\tempo| | | |rary\colburn | | | |dissertation\de| | | |rived\cens_mrg.| | | |sav | | |---------------|---------------| | |Active Dataset |DataSet1 | | |---------------|---------------| | |File Label |CENS_MRG: | | | |Sample & | | | |population | | | |individual | | | |demographics | | | |(WRR) | | |---------------|---------------| | |Filter |<none> | | |---------------|---------------| | |Weight |<none> | | |---------------|---------------| | |Split File |<none> | | |---------------|---------------| | |N of Rows in |1309 | | |Working Data | | | |File | | |-------------|---------------|---------------| |Missing Value|Definition of |User-defined | |Handling |Missing |missing values | | | |are treated as | | | |missing. | | |---------------|---------------| | |Cases Used |Statistics for | | | |each table are | | | |based on all | | | |the cases with | | | |valid data in | | | |the specified | | | |range(s) for | | | |all variables | | | |in each table. | |-------------|---------------|---------------| |Syntax |CROSSTABS | | |/TABLES=ageyear| | |BY sex | | |/FORMAT= AVALUE| | |NOINDEX BOX | | |LABELS TABLES | | |/CELLS= COUNT | | |ROW COLUMN | | |TOTAL . | | | | |-------------|---------------|---------------| |Resources |Elapsed Time |0:00:00.27 | | |---------------|---------------| | |Dimensions |2 | | |Requested | | | |---------------|---------------| | |Cells Available|174876 | |-------------|---------------|---------------| [DataSet1] c:\documents and settings\richard\my documents\temporary\colburn dissertation\derived\cens_mrg.sav Case Processing Summary |---------------|----------------------------------------| | |Cases | | |------------|--------------|------------| | |Valid |Missing |Total | | |----|-------|------|-------|----|-------| | |N |Percent|N |Percent|N |Percent| |---------------|----|-------|------|-------|----|-------| |ageyear Age, |604 |46.1% |705 |53.9% |1309|100.0% | |to lower whole | | | | | | | |year (FOSTAUD3)| | | | | | | |* sex Sex of | | | | | | | |individual | | | | | | | |(FOSTAUD3) | | | | | | | |---------------|----|-------|------|-------|----|-------| ageyear Age, to lower whole year (FOSTAUD3) * sex Sex of individual (FOSTAUD3) Crosstabulation |----------|--|---------------|-------------------------|------| | | | |sex Sex of individual |Total | | | | |(FOSTAUD3) | | | | | |---------------|---------| | | | | |1 Male |2 Female| | |----------|--|---------------|---------------|---------|------| |ageyear |0 |Count |10 |11 |21 | |Age, to | |---------------|---------------|---------|------| |lower | |% within |47.6% |52.4% |100.0%| |whole year| |ageyear Age, | | | | |(FOSTAUD3)| |to lower whole | | | | | | |year (FOSTAUD3)| | | | | | |---------------|---------------|---------|------| | | |% within sex |3.4% |3.5% |3.5% | | | |Sex of | | | | | | |individual | | | | | | |(FOSTAUD3) | | | | | | |---------------|---------------|---------|------| | | |% of Total |1.7% |1.8% |3.5% | | |--|---------------|---------------|---------|------| | |1 |Count |13 |13 |26 | | | |---------------|---------------|---------|------| | | |% within |50.0% |50.0% |100.0%| | | |ageyear Age, | | | | | | |to lower whole | | | | | | |year (FOSTAUD3)| | | | | | |---------------|---------------|---------|------| | | |% within sex |4.4% |4.2% |4.3% | | | |Sex of | | | | | | |individual | | | | | | |(FOSTAUD3) | | | | | | |---------------|---------------|---------|------| | | |% of Total |2.2% |2.2% |4.3% | | |--|---------------|---------------|---------|------| | |2 |Count |17 |14 |31 | | | |---------------|---------------|---------|------| | | |% within |54.8% |45.2% |100.0%| | | |ageyear Age, | | | | | | |to lower whole | | | | | | |year (FOSTAUD3)| | | | | | |---------------|---------------|---------|------| | | |% within sex |5.8% |4.5% |5.1% | | | |Sex of | | | | | | |individual | | | | | | |(FOSTAUD3) | | | | | | |---------------|---------------|---------|------| | | |% of Total |2.8% |2.3% |5.1% | | |--|---------------|---------------|---------|------| | |3 |Count |17 |23 |40 | | | |---------------|---------------|---------|------| | | |% within |42.5% |57.5% |100.0%| | | |ageyear Age, | | | | | | |to lower whole | | | | | | |year (FOSTAUD3)| | | | | | |---------------|---------------|---------|------| | | |% within sex |5.8% |7.4% |6.6% | | | |Sex of | | | | | | |individual | | | | | | |(FOSTAUD3) | | | | | | |---------------|---------------|---------|------| | | |% of Total |2.8% |3.8% |6.6% | | |--|---------------|---------------|---------|------| | |4 |Count |15 |19 |34 | | | |---------------|---------------|---------|------| | | |% within |44.1% |55.9% |100.0%| | | |ageyear Age, | | | | | | |to lower whole | | | | | | |year (FOSTAUD3)| | | | | | |---------------|---------------|---------|------| | | |% within sex |5.1% |6.1% |5.6% | | | |Sex of | | | | | | |individual | | | | | | |(FOSTAUD3) | | | | | | |---------------|---------------|---------|------| | | |% of Total |2.5% |3.1% |5.6% | | |--|---------------|---------------|---------|------| | |5 |Count |19 |11 |30 | | | |---------------|---------------|---------|------| | | |% within |63.3% |36.7% |100.0%| | | |ageyear Age, | | | | | | |to lower whole | | | | | | |year (FOSTAUD3)| | | | | | |---------------|---------------|---------|------| | | |% within sex |6.5% |3.5% |5.0% | | | |Sex of | | | | | | |individual | | | | | | |(FOSTAUD3) | | | | | | |---------------|---------------|---------|------| | | |% of Total |3.1% |1.8% |5.0% | | |--|---------------|---------------|---------|------| | |6 |Count |15 |10 |25 | | | |---------------|---------------|---------|------| | | |% within |60.0% |40.0% |100.0%| | | |ageyear Age, | | | | | | |to lower whole | | | | | | |year (FOSTAUD3)| | | | | | |---------------|---------------|---------|------| | | |% within sex |5.1% |3.2% |4.1% | | | |Sex of | | | | | | |individual | | | | | | |(FOSTAUD3) | | | | | | |---------------|---------------|---------|------| | | |% of Total |2.5% |1.7% |4.1% | | |--|---------------|---------------|---------|------| | |7 |Count |14 |8 |22 | | | |---------------|---------------|---------|------| | | |% within |63.6% |36.4% |100.0%| | | |ageyear Age, | | | | | | |to lower whole | | | | | | |year (FOSTAUD3)| | | | | | |---------------|---------------|---------|------| | | |% within sex |4.8% |2.6% |3.6% | | | |Sex of | | | | | | |individual | | | | | | |(FOSTAUD3) | | | | | | |---------------|---------------|---------|------| | | |% of Total |2.3% |1.3% |3.6% | | |--|---------------|---------------|---------|------| | |8 |Count |11 |12 |23 | | | |---------------|---------------|---------|------| | | |% within |47.8% |52.2% |100.0%| | | |ageyear Age, | | | | | | |to lower whole | | | | | | |year (FOSTAUD3)| | | | | | |---------------|---------------|---------|------| | | |% within sex |3.8% |3.9% |3.8% | | | |Sex of | | | | | | |individual | | | | | | |(FOSTAUD3) | | | | | | |---------------|---------------|---------|------| | | |% of Total |1.8% |2.0% |3.8% | | |--|---------------|---------------|---------|------| | |9 |Count |7 |10 |17 | | | |---------------|---------------|---------|------| | | |% within |41.2% |58.8% |100.0%| | | |ageyear Age, | | | | | | |to lower whole | | | | | | |year (FOSTAUD3)| | | | | | |---------------|---------------|---------|------| | | |% within sex |2.4% |3.2% |2.8% | | | |Sex of | | | | | | |individual | | | | | | |(FOSTAUD3) | | | | | | |---------------|---------------|---------|------| | | |% of Total |1.2% |1.7% |2.8% | | |--|---------------|---------------|---------|------| |
All,
I have a question about what I am thinking of as a type of reliability. The project. Respondents complete a daily diary and record who by name (OPName) they had encounters with make a set of ratings for specific aspects of those encouters (RatingA and RatingB). So the data might look like Resp OPName RatingA RatingB 1 sss 3 7 1 bbb 8 8 1 bbb 2 2 1 ttt 3 6 2 xxx 9 9 ... Etc. After the diary is completed and some time has passed respondents are interviewed with a timeline followback type calendar (and without access to their diary). Respondents name who they had encounters with and make a global rating for the encounters with that person. So the data would look like Resp OPName RatingA RatingB 1 sss 4 6 1 bbb 5 4 2 xxx 6 9 ... Etc. The problem I want to ask about right now is reliability of named persons. The number of unique named person might range from 0 to 6 or 7. In the diary we have encounters nested within persons. But suppose we aggregate by resp and Opname. One way to assess reliability is to put the diary and calendar side by side and compute a percent agreement as (matches/(matches+nomatches)) and then average that over respondents. However, is there a better way? Thanks, Gene Maguin |
In reply to this post by Elaine Prentice-Lane
Call me old fashioned but I strongly agree there were advantages of the
v6 text output and the old TABLES output. I am adept at getting the output I wanted. CTABLES can take considerable effort to accomplish what was easy with TABLES. Of course on the other hand CTABLES can do many things TABLES cannot. I recognize you need something more concrete but I'll have to think more about useful examples to document my assertion. One key advantage of v6 ASCII text output was the ability to quickly search the output of a long complex job with many steps for "ERROR" and "WARNING". Some errors can be quite hard to find in the output of later versions, especially given the difficulty the viewer has with long text logs. The v6 production facility defaulted to output with the same name so it was very efficient to run a large job in the background while working on other things, quickly check the output with an extremely fast text viewer when done, and then start the next job if successful. Over time I find myself becoming more dependent on unique features of later versions and using v6 less. But the newer tools have a price to pay with steep learning curves. The programmer in me would love to take the time to work out creative uses of OMS, CTABLES, Python, etc but I must maintain some balance. I am often far more efficient in getting to substantive findings with a complex data set using outdated tools effectively in v6 than climbing the learning curve of yet another language or tool. On the other hand, in the long run some time spent in strategic learning of new features/tools can have a big payoff. -----Original Message----- From: Peck, Jon [mailto:[hidden email]] Sent: Wednesday, January 17, 2007 5:38 AM Subject: Re: SPSS output - Versions 6 -> 13 or 14 The Draft Viewer, as you probably know, comes closest to the old output format, but it will not reproduce the actual structure of v6 tables, since the output was reorganized after that. What is it exactly that you need about v6 format? If you export the regular Viewer output as XML or use OMS to save XML, you can apply XML tools such as XSLT to transform the output into anything you want, including reorganizing it, much more effectively than applying Perl to plain text. And you could then have the output as plain text, html, or other formats. There are many free XSLT engines available, but you do have to learn a different programming model to use it effectively. HTH, Jon Peck -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Elaine Prentice-Lane Sent: Wednesday, January 17, 2007 7:04 AM To: [hidden email] Subject: [SPSSX-L] SPSS output - Versions 6 -> 13 or 14 We are currently running SPSS for windows - versions 6, 13 and 14. We would prefer to run version 13 or 14, rather than version 6. However, we prefer the version 6 output format for tables. Does anyone have any experience of converting version 13/14 output (tables only) to version 6 format? I have tried exporting in draft format and then post-processing (using Perl) to convert to version 6 table format, but without total success. Is there an easier route? Any help would be much appreciated! Elaine -- ***************************************** Elaine Prentice-Lane ([hidden email]) ISER, University of Essex, Wivenhoe Park, Colchester, Essex CO4 3SQ Tel: +44 (0)1206 873402 Fax: +44 (0)1206 873151 ***************************************** |
I've been resisting this for a while (and maybe I should have continued to
do so) but I would like to expand on this part of Dennis' reply. >>One key advantage of v6 ASCII text output was the ability to quickly search the output of a long complex job with many steps for "ERROR" and "WARNING". Some errors can be quite hard to find in the output of later versions, especially given the difficulty the viewer has with long text logs. I've noticed that while errors appear in the journal file at least some warnings do not. An example, I think, is this. Suppose variable name 'xxx' does not exist, then Frequencies xxx. Generates a warning that appears as an identified element in the whatever the thing on the lefthand pane of the output file is called but does not appear in the log file. Conversely, the statement If (xxx eq 9) yyy=8. Will cause an error listing that appears in a text block of the output and in the journal file but not in the lefthand thing in the output window. Stated tactfully, I think this could be improved greatly by doing two things. 1) printing all warning and errors in the journal file. 2) making error messages appear in the lefthand thing of the output window as its own element and marked in red or some other distinctive and unique color and labeled as such. So why do this. In production mode jobs or in in interactive jobs that have been 'completely' debugged being able to search for either 'error' or 'warning' in the journal is very useful. In interactive mode, having errors identified as such (as warnings are) would be a very helpful visual marker. While I can't believe that this hasn't been suggested before, I'm curious as to why spss has never seen fit to fix this (especially the logging of warnings in the journal file). Gene Maguin |
I have used 2 workarounds.
1) using <edit> <options> <viewer> first set only log, warnings, and notes to <shown> everything else to <hidden> and running the syntax. this provides a smaller output file. after that, set all all options to <shown> and run again. or 2) using <edit> <options> <viewer> set all all options to <shown> before running the syntax. Then in old versions to <file> <print> to CutePDF. In newer versions to <file> <export ><pdf>). Art Kendall Social Research Consultants Gene Maguin wrote: >I've been resisting this for a while (and maybe I should have continued to >do so) but I would like to expand on this part of Dennis' reply. > > > >>>One key advantage of v6 ASCII text output was the ability to quickly >>> >>> >search the output of a long complex job with many steps for "ERROR" and >"WARNING". Some errors can be quite hard to find in the output of later >versions, especially given the difficulty the viewer has with long text >logs. > >I've noticed that while errors appear in the journal file at least some >warnings do not. An example, I think, is this. Suppose variable name 'xxx' >does not exist, then > >Frequencies xxx. > >Generates a warning that appears as an identified element in the whatever >the thing on the lefthand pane of the output file is called but does not >appear in the log file. > >Conversely, the statement > >If (xxx eq 9) yyy=8. > >Will cause an error listing that appears in a text block of the output and >in the journal file but not in the lefthand thing in the output window. > >Stated tactfully, I think this could be improved greatly by doing two >things. >1) printing all warning and errors in the journal file. >2) making error messages appear in the lefthand thing of the output window >as its own element and marked in red or some other distinctive and unique >color and labeled as such. > >So why do this. In production mode jobs or in in interactive jobs that have >been 'completely' debugged being able to search for either 'error' or >'warning' in the journal is very useful. In interactive mode, having errors >identified as such (as warnings are) would be a very helpful visual marker. > >While I can't believe that this hasn't been suggested before, I'm curious as >to why spss has never seen fit to fix this (especially the logging of >warnings in the journal file). > >Gene Maguin > > > >
Art Kendall
Social Research Consultants |
If you want to examine the warnings content separately, one easy way to do this is to catch them all with OMS and write out in a convenient format - perhaps html.
Start the job with OMS /SELECT WARNINGS /DESTINATION FORMAT = HTML /OUTFILE='C:/TEMP/WARNINGS.HTM'. Then just check that after the job is done. You could also include LOGS in the selection to capture message that might appear there. FORMAT=TEXT on /DESTINATION would be another possibility. If you are using programmability, you get the errorlevel as the returned value of the spss.Submit function, so it is easy to check for errors programmatically. HTH, Jon Peck -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Art Kendall Sent: Thursday, January 18, 2007 2:20 PM To: [hidden email] Subject: Re: [SPSSX-L] SPSS output - Versions 6 -> 13 or 14 I have used 2 workarounds. 1) using <edit> <options> <viewer> first set only log, warnings, and notes to <shown> everything else to <hidden> and running the syntax. this provides a smaller output file. after that, set all all options to <shown> and run again. or 2) using <edit> <options> <viewer> set all all options to <shown> before running the syntax. Then in old versions to <file> <print> to CutePDF. In newer versions to <file> <export ><pdf>). Art Kendall Social Research Consultants Gene Maguin wrote: >I've been resisting this for a while (and maybe I should have continued to >do so) but I would like to expand on this part of Dennis' reply. > > > >>>One key advantage of v6 ASCII text output was the ability to quickly >>> >>> >search the output of a long complex job with many steps for "ERROR" and >"WARNING". Some errors can be quite hard to find in the output of later >versions, especially given the difficulty the viewer has with long text >logs. > >I've noticed that while errors appear in the journal file at least some >warnings do not. An example, I think, is this. Suppose variable name 'xxx' >does not exist, then > >Frequencies xxx. > >Generates a warning that appears as an identified element in the whatever >the thing on the lefthand pane of the output file is called but does not >appear in the log file. > >Conversely, the statement > >If (xxx eq 9) yyy=8. > >Will cause an error listing that appears in a text block of the output and >in the journal file but not in the lefthand thing in the output window. > >Stated tactfully, I think this could be improved greatly by doing two >things. >1) printing all warning and errors in the journal file. >2) making error messages appear in the lefthand thing of the output window >as its own element and marked in red or some other distinctive and unique >color and labeled as such. > >So why do this. In production mode jobs or in in interactive jobs that have >been 'completely' debugged being able to search for either 'error' or >'warning' in the journal is very useful. In interactive mode, having errors >identified as such (as warnings are) would be a very helpful visual marker. > >While I can't believe that this hasn't been suggested before, I'm curious as >to why spss has never seen fit to fix this (especially the logging of >warnings in the journal file). > >Gene Maguin > > > > |
Free forum by Nabble | Edit this page |