Matching Assistance
Locked
 
|
I need to be able, in the following file formats, to match for each ID key variable in FILE 2, the housing location and date variables to every row in FILE1. The example of what I need is at the end...suggestions are very much appreciated.
These are the structures of the two files I need to match: File 1 ID Incd# Incddate EventCode 1 1 1/1/09 3 1 1 1/1/09 2 1 2 2/1/09 2 2 3 3/1/09 2 File 2 ID Housing AssignDate 1 Hallb 12/31/08 1 Hallc 1/5/09 1 Halla 1/25/09 2 Hallb 1/15/09 DESIRED Final file ID Incd# Incddate EventCode Housing AssignDate 1 1 1/1/09 3 Hallb 12/31/09 Hallc 1/5/09 Halla 1/25/09 1 1 1/1/09 2 Hallb 12/31/09 Hallc 1/5/09 Halla 1/25/09 1 2 2/1/09 2 Hallb 12/31/09 Hallc 1/5/09 Halla 1/25/09 |
|
|
Cfisher,
There are several things that are not clear about what you want to do. The example shows that you want to get 9 records for id=1. But, you do not show the file 1 info being carried across the six new records file 1 records. I think your result should look like ID Incd# Incddate EventCode Housing AssignDate 1 1 1/1/09 3 Hallb 12/31/09 1 1 1/1/09 3 Hallc 1/5/09 1 1 1/1/09 3 Halla 1/25/09 1 1 1/1/09 2 Hallb 12/31/09 1 1 1/1/09 2 Hallc 1/5/09 1 1 1/1/09 2 Halla 1/25/09 1 2 2/1/09 2 Hallb 12/31/09 1 2 2/1/09 2 Hallc 1/5/09 1 2 2/1/09 2 Halla 1/25/09 Do you disagree? Or, is the way you wrote result a 'shorthand' notation? I think this has to be gotten sorted out first. Gene Magin These are the structures of the two files I need to match: File 1 ID Incd# Incddate EventCode 1 1 1/1/09 3 1 1 1/1/09 2 1 2 2/1/09 2 2 3 3/1/09 2 File 2 ID Housing AssignDate 1 Hallb 12/31/08 1 Hallc 1/5/09 1 Halla 1/25/09 2 Hallb 1/15/09 DESIRED Final file ID Incd# Incddate EventCode Housing AssignDate 1 1 1/1/09 3 Hallb 12/31/09 Hallc 1/5/09 Halla 1/25/09 1 1 1/1/09 2 Hallb 12/31/09 Hallc 1/5/09 Halla 1/25/09 1 2 2/1/09 2 Hallb 12/31/09 Hallc 1/5/09 Halla 1/25/09 -- View this message in context: http://www.nabble.com/Matching-Assistance-tp23259685p23259685.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 |
|
|
Sorry for the confusion Gene, thanks for letting me know.
Yes, I was trying to show what you typed out. The idea being, in essence. For each ID number, there will be multiple admission numbers, and for each ID-Admission# pair, they could be multiple Incd# (incidents). Then, for each admission #, there will be a set of housing location assignment dates. What I want to end up with, is, for each instance where there is the same ID#, the same Admission#, but a different Incd#, i need the entire housing history for that Admission to be matched against that combo as many times as there are housing changes. The need is, I need to know in which housing location a participant was when they had an incident. So, if you see a way to compare each incident date to every housing date for that participant on that admission, without doing a merge like that, that would be helpful too.
|
|
|
Cfisher,
I think you might better think of matching file 1 to file 2 rather than the other way around. I may not be understanding you correctly but I think I would restructure file 2 using casestovars so that you go from > File 2 > ID Housing AssignDate > 1 Hallb 12/31/08 > 1 Hallc 1/5/09 > 1 Halla 1/25/09 > 2 Hallb 1/15/09 To File 2 ID Housing1 Housing2 Housing3 AssignDate1 AssignDate2 AssignDate3 1 Hallb Hallc Halla 12/31/08 1/5/09 1/25/09 2 Hallb 1/15/09 . . Then, treat the restructured file 2 as a 'table' file using the Table subcommand of the Match files command so that a file 2 record is 'spread across' as many file 1 records as have a matching id. Then, there's a couple of ways to do the next part, I think. I'd declare housing1 to housing(n) and assigndate1 to assigndate(n) to be vectors. YOu need to see how many housing or assigndate values that you have. Let that value be NPairs. By the way, I assume that assigndate(i) is a date variable and not a string. Loop #i=1 to npairs. + do if (Incddate ge assign(#i)). + compute IncidentLocation=house(#i). + break. + end if. End loop. This is how I'd start. I haven't tested the code. If you have problems, I work to fix it. Gene Maguin -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of cfishernyc Sent: Monday, April 27, 2009 3:26 PM To: [hidden email] Subject: Re: Matching Assistance Sorry for the confusion Gene, thanks for letting me know. Yes, I was trying to show what you typed out. The idea being, in essence. For each ID number, there will be multiple admission numbers, and for each ID-Admission# pair, they could be multiple Incd# (incidents). Then, for each admission #, there will be a set of housing location assignment dates. What I want to end up with, is, for each instance where there is the same ID#, the same Admission#, but a different Incd#, i need the entire housing history for that Admission to be matched against that combo as many times as there are housing changes. The need is, I need to know in which housing location a participant was when they had an incident. So, if you see a way to compare each incident date to every housing date for that participant on that admission, without doing a merge like that, that would be helpful too. Gene Maguin wrote: > > Cfisher, > > There are several things that are not clear about what you want to do. The > example shows that you want to get 9 records for id=1. But, you do not > show > the file 1 info being carried across the six new records file 1 records. I > think your result should look like > > ID Incd# Incddate EventCode Housing AssignDate > 1 1 1/1/09 3 Hallb 12/31/09 > 1 1 1/1/09 3 Hallc 1/5/09 > 1 1 1/1/09 3 Halla 1/25/09 > 1 1 1/1/09 2 Hallb 12/31/09 > 1 1 1/1/09 2 Hallc 1/5/09 > 1 1 1/1/09 2 Halla 1/25/09 > 1 2 2/1/09 2 Hallb 12/31/09 > 1 2 2/1/09 2 Hallc 1/5/09 > 1 2 2/1/09 2 Halla 1/25/09 > > Do you disagree? Or, is the way you wrote result a 'shorthand' notation? I > think this has to be gotten sorted out first. > > Gene Magin > > > > These are the structures of the two files I need to match: > > File 1 > ID Incd# Incddate EventCode > 1 1 1/1/09 3 > 1 1 1/1/09 2 > 1 2 2/1/09 2 > 2 3 3/1/09 2 > > File 2 > ID Housing AssignDate > 1 Hallb 12/31/08 > 1 Hallc 1/5/09 > 1 Halla 1/25/09 > 2 Hallb 1/15/09 > > > DESIRED Final file > > ID Incd# Incddate EventCode Housing AssignDate > 1 1 1/1/09 3 Hallb 12/31/09 > Hallc 1/5/09 > Halla 1/25/09 > 1 1 1/1/09 2 Hallb 12/31/09 > Hallc 1/5/09 > Halla 1/25/09 > 1 2 2/1/09 2 Hallb 12/31/09 > Hallc 1/5/09 > Halla 1/25/09 > ===================== 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 |
«
Return to SPSSX Discussion
|
1 view|%1 views
| Free forum by Nabble | Edit this page |