Data from R to SPSS Modeler
Locked
 
|
Dear Team,
I have a problem concerning the integration of estimated R-Statistics Data back to SPSS Modeler. FIrst I want to explain what I wanted to do: 1) I do my data preparations in SPSS MOdeler 2) I want to estimate some parameters in R-Statistics 3) These parameters I want to use in the SPSS Modeler Stream again. How I tried to do it: R-Statistics Plug-IN is installed and works correctly. After installing and testing the R-Plug-In I wrote following Code: BEGIN PROGRAM R. spss=read.table(file="C://Users//Backup//Desktop//test.txt", header=TRUE, sep=",") attach(spss) library(survival) surv=Surv(IPT_HIS_REAL) survival.data.weibull=survreg(surv~LL_PER_DAY+KM_COUNT+AGE_DAYS+IPT_SDEV,data=spss, dist="weibull",scale=2.0) summary(survival.data.weibull) coeff=summary(survival.data.weibull)$coefficients[1:4] coeff coeff2=as.data.frame(coeff) coeff2 wert=c("coeff2"," ",0, "F10.7", "scale") dict=spssdictionary.CreateSPSSDictionary(wert) spssdictionary.SetDictionaryToSPSS("results",dict) spssdata.SetDataToSPSS("results", coeff2) spssdictionary.EndDataStep() END PROGRAM. At the end of the code I tried to create a new active dataset which I hoped to save as external Data on my desktop. The Problem is that I´m not sure how to handle this problem. Please could you help me to integrate the parameters into the Modeler?! Thanks in advance. Best regards, Stefan |
|
|
You are almost there. The result-saving
step creates a new Statistics dataset, but it does not save it. After
the END PROGRAM, you need standard Statistics syntax to activate the results
dataset and a SAVE command. This will not make Modeler pick it up
in that stream, since you are presumably using a Statistics Output node,
but it can be picked up in a new stream, and you could use Modeler scripting
to tie the two together.
Other notes. No need for the attach command. Usually the data coming into a Statistics Output node is determined by the Modeler stream, so you might want to use that dataset, which is available via spssdata.GetDataFromSPSS rather than reading an unrelated file. HTH, Jon Peck (no "h") aka Kim Senior Software Engineer, IBM [hidden email] new phone: 720-342-5621 From: Lumar <[hidden email]> To: [hidden email], Date: 11/19/2012 07:35 AM Subject: [SPSSX-L] Data from R to SPSS Modeler Sent by: "SPSSX(r) Discussion" <[hidden email]> Dear Team, I have a problem concerning the integration of estimated R-Statistics Data back to SPSS Modeler. *FIrst I want to explain what I wanted to do:* 1) I do my data preparations in SPSS MOdeler 2) I want to estimate some parameters in R-Statistics 3) These parameters I want to use in the SPSS Modeler Stream again. *How I tried to do it:* R-Statistics Plug-IN is installed and works correctly. After installing and testing the R-Plug-In I wrote following Code: BEGIN PROGRAM R. spss=read.table(file="C://Users//Backup//Desktop//test.txt", header=TRUE, sep=",") attach(spss) library(survival) surv=Surv(IPT_HIS_REAL) survival.data.weibull=survreg(surv~LL_PER_DAY+KM_COUNT+AGE_DAYS+IPT_SDEV,data=spss, dist="weibull",scale=2.0) summary(survival.data.weibull) coeff=summary(survival.data.weibull)$coefficients[1:4] coeff coeff2=as.data.frame(coeff) coeff2 wert=c("coeff2"," ",0, "F10.7", "scale") dict=spssdictionary.CreateSPSSDictionary(wert) spssdictionary.SetDictionaryToSPSS("results",dict) spssdata.SetDataToSPSS("results", coeff2) spssdictionary.EndDataStep() END PROGRAM. At the end of the code I tried to create a new active dataset which I hoped to save as external Data on my desktop. The Problem is that I´m not sure how to handle this problem. Please could you help me to integrate the parameters into the Modeler?! Thanks in advance. Best regards, Stefan -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Data-from-R-to-SPSS-Modeler-tp5716308.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 |
|
|
First, Thank you for your Help! :-)
I added some code for activating and saving the dataset. Until the Activation it works without problem. But after the Save-Code I got an error. My Code: BEGIN PROGRAM R. spss=read.table(file="C://Users//Backup//Desktop//test.txt", header=TRUE, sep=",") attach(spss) library(survival) surv=Surv(IPT_HIS_REAL) survival.data.weibull=survreg(surv~LL_PER_DAY+KM_COUNT+AGE_DAYS+IPT_SDEV,data=spss, dist="weibull",scale=2.0) summary(survival.data.weibull) coeff=summary(survival.data.weibull)$coefficients[1:4] coeff coeff2=as.data.frame(coeff) coeff2 wert=c("coeff2"," ",0, "F10.7", "scale") dict=spssdictionary.CreateSPSSDictionary(wert) spssdictionary.SetDictionaryToSPSS("results",dict) spssdata.SetDataToSPSS("results", coeff2) spssdictionary.EndDataStep() END PROGRAM. DATASET ACTIVATE results SAVE OUTFILE="/data/test567.sav". There must be some error but I couldn´t find it. The Error-Text means following to me: After the equal sign should stand a name but the programm could only find /data/test567.sav. Please could you help me to solve that problem?! Best Regards, Stefan |
|
|
Try this simple program in the Statistics
Output node. It produces a proper results.sav file for me. Perhaps
you need a drive letter in the SAVE command.
data list list/x y. begin data 1 4 2 5 3 6 end data dataset name fred. begin program r. dta = spssdata.GetDataFromSPSS() print(summary(dta)) dict = spssdictionary.CreateSPSSDictionary(c('x','',0,"F8.0", "scale"), c('y', '',0,"F8.0", "scale")) spssdictionary.SetDictionaryToSPSS("results", dict) print(dta) spssdata.SetDataToSPSS("results", dta) spssdictionary.EndDataStep() end program. dataset activate results. save outfile="c:/temp/results.sav". Jon Peck (no "h") aka Kim Senior Software Engineer, IBM [hidden email] new phone: 720-342-5621 From: Lumar <[hidden email]> To: [hidden email], Date: 11/19/2012 08:46 AM Subject: Re: [SPSSX-L] Data from R to SPSS Modeler Sent by: "SPSSX(r) Discussion" <[hidden email]> First, Thank you for your Help! :-) I added some code for activating and saving the dataset. Until the Activation it works without problem. But after the Save-Code I got an error. *My Code:* BEGIN PROGRAM R. spss=read.table(file="C://Users//Backup//Desktop//test.txt", header=TRUE, sep=",") attach(spss) library(survival) surv=Surv(IPT_HIS_REAL) survival.data.weibull=survreg(surv~LL_PER_DAY+KM_COUNT+AGE_DAYS+IPT_SDEV,data=spss, dist="weibull",scale=2.0) summary(survival.data.weibull) coeff=summary(survival.data.weibull)$coefficients[1:4] coeff coeff2=as.data.frame(coeff) coeff2 wert=c("coeff2"," ",0, "F10.7", "scale") dict=spssdictionary.CreateSPSSDictionary(wert) spssdictionary.SetDictionaryToSPSS("results",dict) spssdata.SetDataToSPSS("results", coeff2) spssdictionary.EndDataStep() END PROGRAM. DATASET ACTIVATE results SAVE OUTFILE="/data/test567.sav". There must be some error but I couldn´t find it. The Error-Text means following to me: After the equal sign should stand a name but the programm could only find /data/test567.sav. Please could you help me to solve that problem?! Best Regards, Stefan -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Data-from-R-to-SPSS-Modeler-tp5716308p5716315.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 |
Re: Data from R to SPSS Modeler
|
|
In reply to this post by Lumar
Lumar,
I see that you have already gotten a reply from Jon who knows a lot more about R & spss than I do. You don't offer the text of the error message but it seems to be in one of these two lines. DATASET ACTIVATE results SAVE OUTFILE="/data/test567.sav". Perhaps I'm wrong about the syntax rules in this context but I'll point out that there is a '.' missing after 'results'. Gene Maguin -----Original Message----- From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Lumar Sent: Monday, November 19, 2012 10:30 AM To: [hidden email] Subject: Re: Data from R to SPSS Modeler First, Thank you for your Help! :-) I added some code for activating and saving the dataset. Until the Activation it works without problem. But after the Save-Code I got an error. *My Code:* BEGIN PROGRAM R. spss=read.table(file="C://Users//Backup//Desktop//test.txt", header=TRUE, sep=",") attach(spss) library(survival) surv=Surv(IPT_HIS_REAL) survival.data.weibull=survreg(surv~LL_PER_DAY+KM_COUNT+AGE_DAYS+IPT_SDEV,data=spss, dist="weibull",scale=2.0) summary(survival.data.weibull) coeff=summary(survival.data.weibull)$coefficients[1:4] coeff coeff2=as.data.frame(coeff) coeff2 wert=c("coeff2"," ",0, "F10.7", "scale") dict=spssdictionary.CreateSPSSDictionary(wert) spssdictionary.SetDictionaryToSPSS("results",dict) spssdata.SetDataToSPSS("results", coeff2) spssdictionary.EndDataStep() END PROGRAM. DATASET ACTIVATE results SAVE OUTFILE="/data/test567.sav". There must be some error but I couldn´t find it. The Error-Text means following to me: After the equal sign should stand a name but the programm could only find /data/test567.sav. Please could you help me to solve that problem?! Best Regards, Stefan -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Data-from-R-to-SPSS-Modeler-tp5716308p5716315.html Sent from the SPSSX Discussion mailing list archive at Nabble.com. ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD ===================== To manage your subscription to SPSSX-L, send a message to [hidden email] (not to SPSSX-L), with no body text except the command. To leave the list, send the command SIGNOFF SPSSX-L For a list of commands to manage subscriptions, send the command INFO REFCARD |
Re: Data from R to SPSS Modeler
|
|
In reply to this post by Jon K Peck
It's a little cumbersome (though flexible) to write things like "wert=c("coeff2"," ",0, "F10.7", "scale")"
Isn't it possible to coerce a dataframe into spss data where double/float becomes F10.7/scale, numeric/int becomes F8.0/scale, factor becomes F8.0/nominal and character becomes A-format/nominal? something like: as.savFile(as.date.frame(summary(survival.data.weibull)$coefficients[1:4]))
Regards,
Albert-Jan ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ All right, but apart from the sanitation, the medicine, education, wine, public order, irrigation, roads, a fresh water system, and public health, what have the Romans ever done for us? ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
|
|
|
It would be straightforward (as much as
anything in R is straightforward) to write a function to simplify the dictionary
specification for a new dataset taking a data frame as you describe. I
wouldn't call it savFile, though, since it could only generate a dataset,
not an actual sav file. It would be worth, also, mapping ordered
factors to ordinal level.
Jon Peck (no "h") aka Kim Senior Software Engineer, IBM [hidden email] new phone: 720-342-5621 From: Albert-Jan Roskam <[hidden email]> To: Jon K Peck/Chicago/IBM@IBMUS, "[hidden email]" <[hidden email]>, Date: 11/20/2012 06:29 AM Subject: Re: [SPSSX-L] Data from R to SPSS Modeler It's a little cumbersome (though flexible) to write things like "wert=c("coeff2"," ",0, "F10.7", "scale")" Isn't it possible to coerce a dataframe into spss data where double/float becomes F10.7/scale, numeric/int becomes F8.0/scale, factor becomes F8.0/nominal and character becomes A-format/nominal? something like: as.savFile(as.date.frame(summary(survival.data.weibull)$coefficients[1:4])) Regards, Albert-Jan ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ All right, but apart from the sanitation, the medicine, education, wine, public order, irrigation, roads, a fresh water system, and public health, what have the Romans ever done for us? ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ From: Jon K Peck <[hidden email]> To: [hidden email] Sent: Monday, November 19, 2012 5:29 PM Subject: Re: [SPSSX-L] Data from R to SPSS Modeler Try this simple program in the Statistics Output node. It produces a proper results.sav file for me. Perhaps you need a drive letter in the SAVE command. data list list/x y. begin data 1 4 2 5 3 6 end data dataset name fred. begin program r. dta = spssdata.GetDataFromSPSS() print(summary(dta)) dict = spssdictionary.CreateSPSSDictionary(c('x','',0,"F8.0", "scale"), c('y', '',0,"F8.0", "scale")) spssdictionary.SetDictionaryToSPSS("results", dict) print(dta) spssdata.SetDataToSPSS("results", dta) spssdictionary.EndDataStep() end program. dataset activate results. save outfile="c:/temp/results.sav". Jon Peck (no "h") aka Kim Senior Software Engineer, IBM [hidden email] new phone: 720-342-5621 From: Lumar <[hidden email]> To: [hidden email], Date: 11/19/2012 08:46 AM Subject: Re: [SPSSX-L] Data from R to SPSS Modeler Sent by: "SPSSX(r) Discussion" <[hidden email]> First, Thank you for your Help! :-) I added some code for activating and saving the dataset. Until the Activation it works without problem. But after the Save-Code I got an error. *My Code:* BEGIN PROGRAM R. spss=read.table(file="C://Users//Backup//Desktop//test.txt", header=TRUE, sep=",") attach(spss) library(survival) surv=Surv(IPT_HIS_REAL) survival.data.weibull=survreg(surv~LL_PER_DAY+KM_COUNT+AGE_DAYS+IPT_SDEV,data=spss, dist="weibull",scale=2.0) summary(survival.data.weibull) coeff=summary(survival.data.weibull)$coefficients[1:4] coeff coeff2=as.data.frame(coeff) coeff2 wert=c("coeff2"," ",0, "F10.7", "scale") dict=spssdictionary.CreateSPSSDictionary(wert) spssdictionary.SetDictionaryToSPSS("results",dict) spssdata.SetDataToSPSS("results", coeff2) spssdictionary.EndDataStep() END PROGRAM. DATASET ACTIVATE results SAVE OUTFILE="/data/test567.sav". There must be some error but I couldn´t find it. The Error-Text means following to me: After the equal sign should stand a name but the programm could only find /data/test567.sav. Please could you help me to solve that problem?! Best Regards, Stefan -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Data-from-R-to-SPSS-Modeler-tp5716308p5716315.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 |
Re: Data from R to SPSS Modeler
|
|
"as much as anything in R is straightforward": so True!
begin program r.
options(stringsAsFactors=FALSE) dta <- data.frame(aFactor=factor(c(1, 2, 3)), aFloat=as.double(c(1, 2, 3)), aOrdered=ordered(c(1, 2, 3)), aChar=as.character(c("1", "2", "3"))) width <- max(sapply(dta[sapply(dta, is.character)], nchar), na.rm=TRUE) # max is not really ideal
strVar <- sprintf("A%d", width)
printTypes <- c(factor="F8.0", numeric="F12.2", ordered="F8.0", character=strVar)
varTypes <- c(factor=0, numeric=0, ordered=0, character=width) measLevels <- c(factor="nominal", numeric="scale", ordered="ordinal", character="nominal") dataTypes <- lapply(sapply(dta, class), function(x) x[[1]])
varNames <- names(dataTypes) names(varNames) <- names(varTypes) spssDictInfo <- lapply(dataTypes, function(arg) c(varNames[arg], "''", varTypes[arg][[1]], printTypes[arg][[1]], measLevels[arg][[1]])) dict = spssdictionary.CreateSPSSDictionary( spssDictInfo ) spssdictionary.SetDictionaryToSPSS("results", dict) end program. Regards,
Albert-Jan ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ All right, but apart from the sanitation, the medicine, education, wine, public order, irrigation, roads, a fresh water system, and public health, what have the Romans ever done for us? ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~
|
Administrator
|
I am truly bRamaged after staring at that mess ;-)
How about a few in line comments for the benefit of pedagogical purposes. I am convinced that writing tons of uncommented R code is a built in step to job security. Poke it with a stick? Jello code jumping off the table! B0 + B1*(Free software + expensive consultants - documentation ) + B2 * time =$$$$
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" |
|
|
As R code goes, this is actually pretty
straightforward :-( But I think this was just an example. Some
say that R is easy to learn. My take is that R is just about the
most difficult to learn of any of the dozens of programming languages that
I have experienced.
Jon Peck (no "h") aka Kim Senior Software Engineer, IBM [hidden email] new phone: 720-342-5621 From: David Marso <[hidden email]> To: [hidden email], Date: 11/20/2012 09:31 AM Subject: Re: [SPSSX-L] Data from R to SPSS Modeler Sent by: "SPSSX(r) Discussion" <[hidden email]> I am truly bRamaged after staring at that mess ;-) How about a few in line comments for the benefit of pedagogical purposes. I am convinced that writing tons of uncommented R code is a built in step to job security. Poke it with a stick? Jello code jumping off the table! B0 + B1*(Free software + expensive consultants - documentation ) + B2 * time =$$$$ Albert-Jan Roskam wrote > "as much as anything in R is straightforward": so True! > > begin program r. > options(stringsAsFactors=FALSE) > dta <- data.frame(aFactor=factor(c(1, 2, 3)), aFloat=as.double(c(1, 2, > 3)), aOrdered=ordered(c(1, 2, 3)), aChar=as.character(c("1", "2", "3"))) > > width <- max(sapply(dta[sapply(dta, is.character)], nchar), na.rm=TRUE) # > max is not really ideal > strVar <- sprintf("A%d", width) > printTypes <- c(factor="F8.0", numeric="F12.2", ordered="F8.0", > character=strVar) > varTypes <- c(factor=0, numeric=0, ordered=0, character=width) > measLevels <- c(factor="nominal", numeric="scale", ordered="ordinal", > character="nominal") > dataTypes <- lapply(sapply(dta, class), function(x) x[[1]]) > varNames <- names(dataTypes) > names(varNames) <- names(varTypes) > spssDictInfo <- lapply(dataTypes, function(arg) c(varNames[arg], "''", > varTypes[arg][[1]], printTypes[arg][[1]], measLevels[arg][[1]])) > dict = spssdictionary.CreateSPSSDictionary( spssDictInfo ) > spssdictionary.SetDictionaryToSPSS("results", dict) > end program. > > > Regards, > Albert-Jan > > > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > All right, but apart from the sanitation, the medicine, education, wine, > public order, irrigation, roads, a > fresh water system, and public health, what have the Romans ever done for > us? > ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ > > >>________________________________ >>From: Jon K Peck < > peck@.ibm > > >>To: > SPSSX-L@.UGA > >>Sent: Tuesday, November 20, 2012 3:34 PM >>Subject: Re: [SPSSX-L] Data from R to SPSS Modeler >> >> >>It would be straightforward (as much as anything in R is straightforward) to write a function to simplify the dictionary specification for a new dataset taking a data frame as you describe. I wouldn't call it savFile, though, since it could only generate a dataset, not an actual sav file. It would be worth, also, mapping ordered factors to ordinal level. >> >> >>Jon Peck (no "h") aka Kim >>Senior Software Engineer, IBM >> > peck@.ibm >>new phone: 720-342-5621 >> >> >> >> >>From: Albert-Jan Roskam < > fomcl@ > > >>To: Jon K Peck/Chicago/IBM@IBMUS, " > SPSSX-L@.UGA > " < > SPSSX-L@.UGA > >, >>Date: 11/20/2012 06:29 AM >>Subject: Re: [SPSSX-L] Data from R to SPSS Modeler >> >>>________________________________ >> >> >> >>It's a little cumbersome (though flexible) to write things like "wert=c("coeff2"," ",0, "F10.7", "scale")" >>Isn't it possible to coerce a dataframe into spss data where double/float becomes F10.7/scale, numeric/int becomes F8.0/scale, factor becomes F8.0/nominal and character becomes A-format/nominal? >>something like: as.savFile(as.date.frame(summary(survival.data.weibull)$coefficients[1:4])) >> >>Regards, >>Albert-Jan >> >> >>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >>All right, but apart from the sanitation, the medicine, education, wine, public order, irrigation, roads, a >>fresh water system, and public health, what have the Romans ever done for us? >>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >>From: Jon K Peck < > peck@.ibm > > >>To: > SPSSX-L@.UGA > >>Sent: Monday, November 19, 2012 5:29 PM >>Subject: Re: [SPSSX-L] Data from R to SPSS Modeler >> >>Try this simple program in the Statistics Output node. It produces a proper results.sav file for me. Perhaps you need a drive letter in the SAVE command. >> >>data list list/x y. >>begin data >>1 4 >>2 5 >>3 6 >>end data >>dataset name fred. >>begin program r. >>dta = spssdata.GetDataFromSPSS() >>print(summary(dta)) >>dict = spssdictionary.CreateSPSSDictionary(c('x','',0,"F8.0", "scale"), >>c('y', '',0,"F8.0", "scale")) >>spssdictionary.SetDictionaryToSPSS("results", dict) >>print(dta) >>spssdata.SetDataToSPSS("results", dta) >>spssdictionary.EndDataStep() >>end program. >>dataset activate results. >>save outfile="c:/temp/results.sav". >> >> >>Jon Peck (no "h") aka Kim >>Senior Software Engineer, IBM >> > peck@.ibm >>new phone: 720-342-5621 >> >> >> >> >>From: Lumar < > Stefili@ > > >>To: > SPSSX-L@.uga > , >>Date: 11/19/2012 08:46 AM >>Subject: Re: [SPSSX-L] Data from R to SPSS Modeler >>Sent by: "SPSSX(r) Discussion" < > SPSSX-L@.uga > > >> >> >> >>First, Thank you for your Help! :-) >> >>I added some code for activating and saving the dataset. Until the >>Activation it works without problem. But after the Save-Code I got an error. >> >>*My Code:* >> >>BEGIN PROGRAM R. >>spss=read.table(file="C://Users//Backup//Desktop//test.txt", header=TRUE, >>sep=",") >> >> >>attach(spss) >>library(survival) >>surv=Surv(IPT_HIS_REAL) >> >>survival.data.weibull=survreg(surv~LL_PER_DAY+KM_COUNT+AGE_DAYS+IPT_SDEV,data=spss, >>dist="weibull",scale=2.0) >>summary(survival.data.weibull) >> >>coeff=summary(survival.data.weibull)$coefficients[1:4] >>coeff >>coeff2=as.data.frame(coeff) >>coeff2 >> >>wert=c("coeff2"," ",0, "F10.7", "scale") >>dict=spssdictionary.CreateSPSSDictionary(wert) >>spssdictionary.SetDictionaryToSPSS("results",dict) >>spssdata.SetDataToSPSS("results", coeff2) >>spssdictionary.EndDataStep() >> >>END PROGRAM. >>DATASET ACTIVATE results >>SAVE OUTFILE="/data/test567.sav". >> >>There must be some error but I couldn´t find it. The Error-Text means >>following to me: >>After the equal sign should stand a name but the programm could only find >>/data/test567.sav. >> >> >>Please could you help me to solve that problem?! >> >>Best Regards, >>Stefan >> >> >> >>-- >>View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Data-from-R-to-SPSS-Modeler-tp5716308p5716315.html >>Sent from the SPSSX Discussion mailing list archive at Nabble.com. >> >>===================== >>To manage your subscription to SPSSX-L, send a message to >> > LISTSERV@.UGA > (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 >> >> >> >> >> >> >> ----- Please reply to the list and not to my personal email. Those desiring my consulting or training services please feel free to email me. -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Data-from-R-to-SPSS-Modeler-tp5716308p5716350.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 |
|
Administrator
|
"As R code goes, this is actually pretty straightforward :-( "..
I was afraid that was the case! So far my experience as well. Currently creating SPSS code to implement methods in a couple of articles. Also have R code which would hopefully have illuminated some of the trickier parts of the equations (no such luck). Just for my edification Jon: What is it about R that makes it the 'most difficult to learn' in your experience? ---
Please reply to the list and not to my personal email.
Those desiring my consulting or training services please feel free to email me. --- "Nolite dare sanctum canibus neque mittatis margaritas vestras ante porcos ne forte conculcent eas pedibus suis." Cum es damnatorum possederunt porcos iens ut salire off sanguinum cliff in abyssum?" |
Re: Data from R to SPSS Modeler
|
|
In reply to this post by David Marso
>I am truly bRamaged after staring at that mess ;-)
>How about a few in line comments for the benefit of pedagogical purposes. >I am convinced that writing tons of uncommented R code is a built in step to >job security. >Poke it with a stick? Jello code jumping off the table! >B0 + B1*(Free software + expensive consultants - documentation ) + B2 * time >=$$$$ Hahaha, sorry about that, but don't shoot the messenger! R is a language of which one forgets the details after about 15mins not using it. With lots of prentheses, curly braces and square brackets, and most defaults not what you'd like them to be. The *ply functions may be a bit awkward in the beginning, but they're really handy once you "get" them. But it can be so hard not to think in terms of "for" loops! One option() that is missing in R is: options(dontCoerceEverything=True). I hate it that dataframes (or other data types) are coerced into lists, matrices, get additional attributes, etc. etc. The code below is just a mapping between R datatypes and SPSS data/printtypes (spss has just two datatypes, of course) and measurement levels. double --> F12.2, scale numeric (int) --> F8, scale factor --> F8, nominal ordered --> F8. ordinal character --> A<width>, nominal I wrote the code because I found it awkward to specify the arguments of spssdictionary.SetDictionaryToSPSS manually. Too many lines of code ;-)))) >Albert-Jan Roskam wrote >> "as much as anything in R is straightforward": so True! >>  >> begin program r. >> options(stringsAsFactors=FALSE) >> dta <- data.frame(aFactor=factor(c(1, 2, 3)), aFloat=as.double(c(1, 2, >> 3)), aOrdered=ordered(c(1, 2, 3)), aChar=as.character(c("1", "2", "3"))) >> >> width <- max(sapply(dta[sapply(dta, is.character)], nchar), na.rm=TRUE) # >> max is not really ideal >> strVar <- sprintf("A%d", width) >> printTypes <- c(factor="F8.0", numeric="F12.2", ordered="F8.0", >> character=strVar) >> varTypes <- c(factor=0, numeric=0, ordered=0, character=width) >> measLevels <- c(factor="nominal", numeric="scale", ordered="ordinal", >> character="nominal") >> dataTypes <- lapply(sapply(dta, class), function(x) x[[1]]) >> varNames <- names(dataTypes) >> names(varNames) <- names(varTypes) >> spssDictInfo <- lapply(dataTypes, function(arg) c(varNames[arg], "''", >> varTypes[arg][[1]], printTypes[arg][[1]], measLevels[arg][[1]])) >> dict = spssdictionary.CreateSPSSDictionary( spssDictInfo ) >> spssdictionary.SetDictionaryToSPSS("results", dict) >> end program. >> >> >> Regards, >> Albert-Jan >> >> >> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >> All right, but apart from the sanitation, the medicine, education, wine, >> public order, irrigation, roads, a >> fresh water system, and public health, what have the Romans ever done for >> us? >> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >> >> >>>________________________________ >>>From: Jon K Peck < > >> peck@.ibm > >> > >>>To: > >> SPSSX-L@.UGA > >> >>>Sent: Tuesday, November 20, 2012 3:34 PM >>>Subject: Re: [SPSSX-L] Data from R to SPSS Modeler >>> >>> >>>It would be straightforward (as much as anything in R is straightforward) >to write a function to simplify the dictionary specification for a new >dataset taking a data frame as you describe.  I wouldn't call it savFile, >though, since it could only generate a dataset, not an actual sav file.  It >would be worth, also, mapping ordered factors to ordinal level. >>> >>> >>>Jon Peck (no "h") aka Kim >>>Senior Software Engineer, IBM >>> > >> peck@.ibm > >>>new phone: 720-342-5621 >>> >>> >>> >>> >>>From:     Albert-Jan Roskam < > >> fomcl@ > >> > >>>To:     Jon K Peck/Chicago/IBM@IBMUS, " > >> SPSSX-L@.UGA > >> " < > >> SPSSX-L@.UGA > >> >, >>>Date:     11/20/2012 06:29 AM >>>Subject:     Re: [SPSSX-L] Data from R to SPSS Modeler >>> >>>>________________________________ >>> >>> >>> >>>It's a little cumbersome (though flexible) to write things like > "wert=c("coeff2"," ",0, "F10.7", "scale")" >>>Isn't it possible to coerce a dataframe into spss data where double/float >becomes F10.7/scale, numeric/int becomes F8.0/scale, factor becomes >F8.0/nominal and character becomes A-format/nominal? >>>something like: >as.savFile(as.date.frame(summary(survival.data.weibull)$coefficients[1:4])) >>> >>>Regards, >>>Albert-Jan >>> >>> >>>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >>>All right, but apart from the sanitation, the medicine, education, wine, >public order, irrigation, roads, a >>>fresh water system, and public health, what have the Romans ever done for >us? >>>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >>>From: Jon K Peck < > >> peck@.ibm > >> > >>>To: > >> SPSSX-L@.UGA > >> >>>Sent: Monday, November 19, 2012 5:29 PM >>>Subject: Re: [SPSSX-L] Data from R to SPSS Modeler >>> >>>Try this simple program in the Statistics Output node.  It produces a >proper results.sav file for me.  Perhaps you need a drive letter in the SAVE >command. >>> >>>data list list/x y. >>>begin data >>>1 4 >>>2 5 >>>3 6 >>>end data >>>dataset name fred. >>>begin program r. >>>dta = spssdata.GetDataFromSPSS() >>>print(summary(dta)) >>>dict = spssdictionary.CreateSPSSDictionary(c('x','',0,"F8.0", "scale"), >>>c('y', '',0,"F8.0", "scale")) >>>spssdictionary.SetDictionaryToSPSS("results", dict) >>>print(dta) >>>spssdata.SetDataToSPSS("results", dta) >>>spssdictionary.EndDataStep() >>>end program. >>>dataset activate results. >>>save outfile="c:/temp/results.sav". >>> >>> >>>Jon Peck (no "h") aka Kim >>>Senior Software Engineer, IBM >>> > >> peck@.ibm > >>>new phone: 720-342-5621 >>> >>> >>> >>> >>>From:     Lumar < > >> Stefili@ > >> > >>>To:     > >> SPSSX-L@.uga > >> , >>>Date:     11/19/2012 08:46 AM >>>Subject:     Re: [SPSSX-L] Data from R to SPSS Modeler >>>Sent by:     "SPSSX(r) Discussion" < > >> SPSSX-L@.uga > >> > >>> >>> >>> >>>First, Thank you for your Help! :-) >>> >>>I added some code for activating and saving the dataset. Until the >>>Activation it works without problem. But after the Save-Code I got an >error. >>> >>>*My Code:* >>> >>>BEGIN PROGRAM R. >>>spss=read.table(file="C://Users//Backup//Desktop//test.txt", header=TRUE, >>>sep=",") >>> >>> >>>attach(spss) >>>library(survival) >>>surv=Surv(IPT_HIS_REAL) >>> >>>survival.data.weibull=survreg(surv~LL_PER_DAY+KM_COUNT+AGE_DAYS+IPT_SDEV,data=spss, >>>dist="weibull",scale=2.0) >>>summary(survival.data.weibull) >>> >>>coeff=summary(survival.data.weibull)$coefficients[1:4] >>>coeff >>>coeff2=as.data.frame(coeff) >>>coeff2 >>> >>>wert=c("coeff2"," ",0, "F10.7", "scale") >>>dict=spssdictionary.CreateSPSSDictionary(wert) >>>spssdictionary.SetDictionaryToSPSS("results",dict) >>>spssdata.SetDataToSPSS("results", coeff2) >>>spssdictionary.EndDataStep() >>> >>>END PROGRAM. >>>DATASET ACTIVATE results >>>SAVE OUTFILE="/data/test567.sav". >>> >>>There must be some error but I couldn´t find it. The Error-Text means >>>following to me: >>>After the equal sign should stand a name but the programm could only find >>>/data/test567.sav. >>> >>> >>>Please could you help me to solve that problem?! >>> >>>Best Regards, >>>Stefan >>> >>> >>> >>>-- >>>View this message in context: >http://spssx-discussion.1045642.n5.nabble.com/Data-from-R-to-SPSS-Modeler-tp5716308p5716315.html >>>Sent from the SPSSX Discussion mailing list archive at Nabble.com. >>> >>>===================== >>>To manage your subscription to SPSSX-L, send a message to >>> > >> LISTSERV@.UGA > >> (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 >>> >>> >>> >>> >>> >>> >>> > > > > > >----- >Please reply to the list and not to my personal email. >Those desiring my consulting or training services please feel free to email me. >-- >View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Data-from-R-to-SPSS-Modeler-tp5716308p5716350.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 |
|
|
In reply to this post by David Marso
R has a huge number of packages (a problem
in itself, sometimes), and powerful tools for implementing statistical
and graphical algorithms. Those libraries are probably the biggest
benefit for SPSS users. But as a language, it's pretty incoherent
- somewhat object oriented, somewhat not, notationally inconsistent, (those
dots in names do not indicate attributes of objects). Syntax is C
like but many things are different. Scoping of data and functions
is way too complex and does not do a good job of information hiding or
abstraction.
There are thousands of functions, so there are usually twelve different ways to do something where one general method would do - okay, maybe two. For example, it's hard to remember which of the workhorse apply functions to use: apply, lapply, tapply, sapply, mapply, or eapply. They all do similar but not identical things. Many functions have quite complex argument sets, albeit often with defaults, that make them hard to use. Because many ordinary operations are vectorized, simple things like pasting strings together can be hard to figure out. And package authors - anyone can contribute - are often inconsistent from one to another. To top it off, R documentation is written "by experts for experts". Debugging is a trial. I spent several hours this morning figuring out how to do something that should be simple and could be done trivially in Python, but I waded through at least 30 R functions and a bunch of Googling figuring out most of what I needed. I've copied an extract from the doc for the do.call function below as a somewhat more difficult than average example of the documentation. Getting that off my chest! If R were being designed today by skilled language designers, many things would be different, I believe, both in the design and in the internals, but for better or worse, those libraries and the power of the language mean that for statistical development, it's still a great resource. And if IBM or someone else wraps R functionality in a Statistics extension command, you get the power without the pain. ------------------------- do.call constructs and executes a function call from a name or a function and a list of arguments to be passed to it. Usage
Details If quote is FALSE, the default, then the arguments are evaluated (in the calling environment, not in envir). If quote is TRUE then each argument is quoted (see quote) so that the effect of argument evaluation is to remove the quotes – leaving the original arguments unevaluated when the call is constructed. The behavior of some functions, such as substitute,
will not be the same for functions evaluated using do.call
as if they were evaluated from the interpreter. The precise semantics are
currently undefined and subject to change.
"As R code goes, this is actually pretty straightforward :-( ".. I was afraid that was the case! So far my experience as well. Currently creating SPSS code to implement methods in a couple of articles. Also have R code which would hopefully have illuminated some of the trickier parts of the equations (no such luck). Just for my edification Jon: What is it about R that makes it the 'most difficult to learn' in your experience? --- Jon K Peck wrote > As R code goes, this is actually pretty straightforward :-( But I think > this was just an example. Some say that R is easy to learn. My take is > that R is just about the most difficult to learn of any of the dozens of > programming languages that I have experienced. > > > Jon Peck (no "h") aka Kim > Senior Software Engineer, IBM > peck@.ibm > new phone: 720-342-5621 > > > > > From: David Marso < > david.marso@ > > > To: > SPSSX-L@.uga > , > Date: 11/20/2012 09:31 AM > Subject: Re: [SPSSX-L] Data from R to SPSS Modeler > Sent by: "SPSSX(r) Discussion" < > SPSSX-L@.uga > > > > > > I am truly bRamaged after staring at that mess ;-) > How about a few in line comments for the benefit of pedagogical purposes. > I am convinced that writing tons of uncommented R code is a built in step > to > job security. > Poke it with a stick? Jello code jumping off the table! > B0 + B1*(Free software + expensive consultants - documentation ) + B2 * > time > =$$$$ > > > > Albert-Jan Roskam wrote >> "as much as anything in R is straightforward": so True! >> >> begin program r. >> options(stringsAsFactors=FALSE) >> dta <- data.frame(aFactor=factor(c(1, 2, 3)), aFloat=as.double(c(1, 2, >> 3)), aOrdered=ordered(c(1, 2, 3)), aChar=as.character(c("1", "2", "3"))) >> >> width <- max(sapply(dta[sapply(dta, is.character)], nchar), na.rm=TRUE) > # >> max is not really ideal >> strVar <- sprintf("A%d", width) >> printTypes <- c(factor="F8.0", numeric="F12.2", ordered="F8.0", >> character=strVar) >> varTypes <- c(factor=0, numeric=0, ordered=0, character=width) >> measLevels <- c(factor="nominal", numeric="scale", ordered="ordinal", >> character="nominal") >> dataTypes <- lapply(sapply(dta, class), function(x) x[[1]]) >> varNames <- names(dataTypes) >> names(varNames) <- names(varTypes) >> spssDictInfo <- lapply(dataTypes, function(arg) c(varNames[arg], "''", >> varTypes[arg][[1]], printTypes[arg][[1]], measLevels[arg][[1]])) >> dict = spssdictionary.CreateSPSSDictionary( spssDictInfo ) >> spssdictionary.SetDictionaryToSPSS("results", dict) >> end program. >> >> >> Regards, >> Albert-Jan >> >> >> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >> All right, but apart from the sanitation, the medicine, education, wine, >> public order, irrigation, roads, a >> fresh water system, and public health, what have the Romans ever done > for >> us? >> ~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >> >> >>>________________________________ >>>From: Jon K Peck < > >> peck@.ibm > >> > >>>To: > >> SPSSX-L@.UGA > >> >>>Sent: Tuesday, November 20, 2012 3:34 PM >>>Subject: Re: [SPSSX-L] Data from R to SPSS Modeler >>> >>> >>>It would be straightforward (as much as anything in R is > straightforward) > to write a function to simplify the dictionary specification for a new > dataset taking a data frame as you describe. I wouldn't call it savFile, > though, since it could only generate a dataset, not an actual sav file. It > would be worth, also, mapping ordered factors to ordinal level. >>> >>> >>>Jon Peck (no "h") aka Kim >>>Senior Software Engineer, IBM >>> > >> peck@.ibm > >>>new phone: 720-342-5621 >>> >>> >>> >>> >>>From: Albert-Jan Roskam < > >> fomcl@ > >> > >>>To: Jon K Peck/Chicago/IBM@IBMUS, " > >> SPSSX-L@.UGA > >> " < > >> SPSSX-L@.UGA > >> >, >>>Date: 11/20/2012 06:29 AM >>>Subject: Re: [SPSSX-L] Data from R to SPSS Modeler >>> >>>>________________________________ >>> >>> >>> >>>It's a little cumbersome (though flexible) to write things like > "wert=c("coeff2"," ",0, "F10.7", "scale")" >>>Isn't it possible to coerce a dataframe into spss data where > double/float > becomes F10.7/scale, numeric/int becomes F8.0/scale, factor becomes > F8.0/nominal and character becomes A-format/nominal? >>>something like: > as.savFile(as.date.frame(summary(survival.data.weibull)$coefficients[1:4])) >>> >>>Regards, >>>Albert-Jan >>> >>> >>>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >>>All right, but apart from the sanitation, the medicine, education, wine, > public order, irrigation, roads, a >>>fresh water system, and public health, what have the Romans ever done > for > us? >>>~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~~ >>>From: Jon K Peck < > >> peck@.ibm > >> > >>>To: > >> SPSSX-L@.UGA > >> >>>Sent: Monday, November 19, 2012 5:29 PM >>>Subject: Re: [SPSSX-L] Data from R to SPSS Modeler >>> >>>Try this simple program in the Statistics Output node. It produces a > proper results.sav file for me. Perhaps you need a drive letter in the > SAVE > command. >>> >>>data list list/x y. >>>begin data >>>1 4 >>>2 5 >>>3 6 >>>end data >>>dataset name fred. >>>begin program r. >>>dta = spssdata.GetDataFromSPSS() >>>print(summary(dta)) >>>dict = spssdictionary.CreateSPSSDictionary(c('x','',0,"F8.0", "scale"), >>>c('y', '',0,"F8.0", "scale")) >>>spssdictionary.SetDictionaryToSPSS("results", dict) >>>print(dta) >>>spssdata.SetDataToSPSS("results", dta) >>>spssdictionary.EndDataStep() >>>end program. >>>dataset activate results. >>>save outfile="c:/temp/results.sav". >>> >>> >>>Jon Peck (no "h") aka Kim >>>Senior Software Engineer, IBM >>> > >> peck@.ibm > >>>new phone: 720-342-5621 >>> >>> >>> >>> >>>From: Lumar < > >> Stefili@ > >> > >>>To: > >> SPSSX-L@.uga > >> , >>>Date: 11/19/2012 08:46 AM >>>Subject: Re: [SPSSX-L] Data from R to SPSS Modeler >>>Sent by: "SPSSX(r) Discussion" < > >> SPSSX-L@.uga > >> > >>> >>> >>> >>>First, Thank you for your Help! :-) >>> >>>I added some code for activating and saving the dataset. Until the >>>Activation it works without problem. But after the Save-Code I got an > error. >>> >>>*My Code:* >>> >>>BEGIN PROGRAM R. >>>spss=read.table(file="C://Users//Backup//Desktop//test.txt", > header=TRUE, >>>sep=",") >>> >>> >>>attach(spss) >>>library(survival) >>>surv=Surv(IPT_HIS_REAL) >>> >>>survival.data.weibull=survreg(surv~LL_PER_DAY+KM_COUNT+AGE_DAYS+IPT_SDEV,data=spss, >>>dist="weibull",scale=2.0) >>>summary(survival.data.weibull) >>> >>>coeff=summary(survival.data.weibull)$coefficients[1:4] >>>coeff >>>coeff2=as.data.frame(coeff) >>>coeff2 >>> >>>wert=c("coeff2"," ",0, "F10.7", "scale") >>>dict=spssdictionary.CreateSPSSDictionary(wert) >>>spssdictionary.SetDictionaryToSPSS("results",dict) >>>spssdata.SetDataToSPSS("results", coeff2) >>>spssdictionary.EndDataStep() >>> >>>END PROGRAM. >>>DATASET ACTIVATE results >>>SAVE OUTFILE="/data/test567.sav". >>> >>>There must be some error but I couldn´t find it. The Error-Text means >>>following to me: >>>After the equal sign should stand a name but the programm could only > find >>>/data/test567.sav. >>> >>> >>>Please could you help me to solve that problem?! >>> >>>Best Regards, >>>Stefan >>> >>> >>> >>>-- >>>View this message in context: > http://spssx-discussion.1045642.n5.nabble.com/Data-from-R-to-SPSS-Modeler-tp5716308p5716315.html > >>>Sent from the SPSSX Discussion mailing list archive at Nabble.com. >>> >>>===================== >>>To manage your subscription to SPSSX-L, send a message to >>> > >> LISTSERV@.UGA > >> (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 >>> >>> >>> >>> >>> >>> >>> > > > > > > ----- > Please reply to the list and not to my personal email. > Those desiring my consulting or training services please feel free to > email me. > -- > View this message in context: > http://spssx-discussion.1045642.n5.nabble.com/Data-from-R-to-SPSS-Modeler-tp5716308p5716350.html > > Sent from the SPSSX Discussion mailing list archive at Nabble.com. > > ===================== > To manage your subscription to SPSSX-L, send a message to > LISTSERV@.UGA > (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 ----- Please reply to the list and not to my personal email. Those desiring my consulting or training services please feel free to email me. -- View this message in context: http://spssx-discussion.1045642.n5.nabble.com/Data-from-R-to-SPSS-Modeler-tp5716308p5716359.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 |
Re: Data from R to SPSS Modeler
|
|
Everyone: I’ve been looking at this thread with interest but I did not intend to reply until the value of R was questioned. I’ve used SPSS for more than 20 years and I plan to continue using it in our office since I am responsible for an extensive set of analyses that are based on our long history with SPSS and I would not want to even think of what we would need to do to use another program. However, for new projects where we have no history I increasingly use R and as you can imagine, I do not agree with Dr. Peck’s comments (below) about the difficulty of R. The SPSS developers would do well to look at the many graphics options available through R that simply do not exist in SPSS. If you wish to look at R-based syntax that is easy to follow and nicely documented then review the text at http://www.springer.com/statistics/social+sciences+%26+law/book/978-1-4614-2133-7, specifically: No – I’m not touting my book. I just know this resource better than others. Best wishes and Happy Thanksgiving. Tom ---------- Thomas W. MacFarland, Ed.D. Senior Research Associate; Institutional Effectiveness and Associate Professor Nova Southeastern University Voice 954-262-5395 [hidden email] From: SPSSX(r) Discussion [mailto:[hidden email]] On Behalf Of Jon K Peck R has a huge number of packages (a problem in itself, sometimes), and powerful tools for implementing statistical and graphical algorithms. Those libraries are probably the biggest benefit for SPSS users. But as a language, it's pretty incoherent - somewhat object oriented, somewhat not, notationally inconsistent, (those dots in names do not indicate attributes of objects). Syntax is C like but many things are different. Scoping of data and functions is way too complex and does not do a good job of information hiding or abstraction. Usage
Details If quote is FALSE, the default, then the arguments are evaluated (in the calling environment, not in envir). If quote is TRUE then each argument is quoted (see quote) so that the effect of argument evaluation is to remove the quotes – leaving the original arguments unevaluated when the call is constructed. The behavior of some functions, such as substitute, will not be the same for functions evaluated using do.call as if they were evaluated from the interpreter. The precise semantics are currently undefined and subject to change.
|
|
|
In reply to this post by Jon K Peck
sorry for interrupting this discussion but I want to thank you all for your help.
My Code seems to be correct now! Only the "." after results was missing. Thank you very much! |
«
Return to SPSSX Discussion
|
1 view|%1 views
| Free forum by Nabble | Edit this page |