Re: Desktop Reporter - Excel file formatting

[Dennis Deck]

Not familiar with Desktop Reporter but suspect this is a more generic
problem typical in using spreadsheet as data source.

 - One issue with Excel is that a column of identifiers may default to
numeric rather than text (or vice versa) in one file based on what it
sees in that column. Given that source was different it is quite
possible that an ID variable (or other variable) was handled differently
by the two sources.

 - The DBNull error message suggests to me that there might have been
empty cells - if that is true of the matching variable then you may need
either drop the record or enter a value.

Not sure what Desktop Reporter is.  May want to try without that.
Another layer of software could well be creating problems.

Dennis Deck, PhD
RMC Research Corporation
111 SW Columbia Street, Suite 1200
Portland, Oregon 97201-5843
voice: 503-223-8248 x715
voice: 800-788-1887 x715
fax:  [phone]
[hidden email]


-----Original Message-----
From: Tierney, Mary Lou [mailto:[hidden email]]
Sent: Wednesday, June 11, 2008 6:13 AM
Subject: FW: Desktop Reporter - Excel file formatting



Hello,



I am trying to merge data (Merge variables from one other data set) from
two Excel documents but I keep getting a failed merge message:
object cannot be cast from DBNull to other types.



I am a new Desktop reporter user - can anyone help me understand what
Excel formatting issues will cause a failed merge, i.e. blank cells,
etc.



One of the Excel files is also a 'Categories Export" (dichotomies) from
SPSS Text Analysis - is anyone aware of any issues with exported Excel
files from Text Analysis that get imported into Desktop reporter?



Looking for any insights - SPSS Tech Support has been abysmal.







Mary Lou Tierney

Enterprise Architecture Planning and Innovation

The MITRE Corporation

p  781.271.8862

e  [hidden email]

=====================
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

[Sam Winstanley]

I know that this is perhaps a couple of months too late.

I have a strong hunch your best plan is to convert the data from excel to something else before you do the merge. The most simple way is to just open the Excel file in reporter then do File->Export Data to get it out. I would suggest exporting Dimensions Data File.

Regarding the handling of Dichotomy sets that come from excel, thats a tough one as we stand (with Reporter v5.5) however if you can get Category sets out this walkthrough might help you:
http://www.surveyxtreme.com/?p=74

If you can get the data out of STAfS (SPSS Text Analysis for Surveys) in another format you will be in better shape too.

Put simply, Excel data is just a bit weak in metadata to handle in Reporter without some initial work setting it up. Most of the time its better to use somethign with better descriptive capabilities (such as Dimensions Data File, or SPSS file).

[Suvash Dhakal]

Hey Sam/and others,

I was playing with SPSS multiple response table. After creating a
multiple response sets, I saved the file and exported to Desktop
reporter. After that, it was easier to crosstab the Multiple response
question in Desktop reporter. It also saves lot of time.

But now I have a different question for you..a tough one for me.! How to
create a grip table for rating/scaling questions (rating 1-5) when
different statements for a same variable are in different column. What
DR is doing right now is treating different statements, let's say 1a,
1b, 1c, 1d (all in different column) as separate questions. What I am
looking is trying merge all these statement into a single table so that
if we want top box (ratings 4, 5), we just have a single table..Any
ideas would be immensely useful for novice DR user like me and others
trying to figure this out!

-Suvash


-----Original Message-----
From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of
Sam Winstanley
Sent: Wednesday, August 13, 2008 11:04 AM
To: [hidden email]
Subject: Re: Desktop Reporter - Excel file formatting

I know that this is perhaps a couple of months too late.

I have a strong hunch your best plan is to convert the data from excel
to
something else before you do the merge. The most simple way is to just
open
the Excel file in reporter then do File->Export Data to get it out. I
would
suggest exporting Dimensions Data File.

Regarding the handling of Dichotomy sets that come from excel, thats a
tough
one as we stand (with Reporter v5.5) however if you can get Category
sets
out this walkthrough might help you:
http://www.surveyxtreme.com/?p=74 http://www.surveyxtreme.com/?p=74

If you can get the data out of STAfS (SPSS Text Analysis for Surveys) in
another format you will be in better shape too.

Put simply, Excel data is just a bit weak in metadata to handle in
Reporter
without some initial work setting it up. Most of the time its better to
use
somethign with better descriptive capabilities (such as Dimensions Data
File, or SPSS file).




--
View this message in context:
http://www.nabble.com/Re%3A-Desktop-Reporter---Excel-file-formatting-tp1
7809198p18964905.html
Sent from the SPSSX Discussion mailing list archive at Nabble.com.

=====================
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

=====================
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

[Sam Winstanley]

Hi,

Although Desktop Reporter has lots of nice features for tabulting grid and loop variables, at the present time (Dimensions version 5.5) there is no way to create a New Grid variable within the confines of Desktop Reporter.

For instance, if you got this kind of data from mrInterview (or actually a few other sources which supports heirarchical data and grids) then you would be completely able to create this kind of data in the way you suggest.

This is the kind of thing I would not be surprised to see SPSS add in the future.

It's certainly possible to convert non-gridded data into gridded data using mrStudio and scripting, or by building programs which work directly against the Dimension's data. However not everything you can do by the technical API's is currently available as a Desktop Reporter feature, and I suspect this might always be the case.

Thanks
Sam