R help - Apr 2010 - Help with replacement of certain values in dataset with SAS code equivalent

I am new to R and have tried for a good while to figure out how to code this
in R.  The dataset below:

 


FTIStandKey

State

County

FTITract

CoverType

Ver_CT

V_Origin

V_SpGrp

	


NAH6005-001

Texas

Jasper

NAH6005

PPLB-2000-U

PPLB-2000-U

P

P

	


NAH6005-002

Texas

Jasper

NAH6005

NHHX-1950-O

NHHX-1950-I

N

H

	


NAH6253-001

Texas

Tyler

NAH6253

PPLB-2001-U

PPLB-2001-U

P

P

	


NAH6253-002

Texas

Tyler

NAH6253

PPLB-1993-U

PPLB-1993-U

P

P

	


NAH6253-003

Texas

Tyler

NAH6253

PPLB-2003-U

PPLB-2003-U

P

P

	


NAH6253-005

Texas

Tyler

NAH6253

XSOP-9999-C

XSOP-9999-C

X

S

	


NAH6253-008

Texas

Tyler

NAH6253

NPLB-1975-O

NPLB-1975-O

N

P

	


 

The SAS code for what I want to do is:

 

/*

 if FTIStandKey=" NAH6253-003" then do;

   CoverType=" XSOP-9999-C ";

   V_SpGrp="T";

   V_Origin="T";

 end;

 

What I want to do is to be able to select certain FTIStandKey's and change
values in the dataset.  For example, I would like to select by
FTIStandKey=="NAH6253-003", and change the entries for
CoverType="
XSOP-9999-C " and change the V_SpGrp=="T", and also change
V_Origin=="T".  I
only want to change the record for the one FTIStandKey.  All the other
records need to remain the exact same.  I have spent a good bit of time
trying to figure this out but with no success.   Any help would be greatly
appreciated.

 

The only thing I could find even close to what I wanted to do is listed
below:

 

# R Program for Multiple Conditional Transformations.

# Read the file into a data frame and print it.

load(file="c:\\mydata.Rdata")

print(mydata)

 

# Use column bind to add two new columns to mydata.

# Not necessary for this example, but handy to know.

 

mydata <- cbind(mydata, score1 = 0, score2 = 0)

 

attach(mydata)

mydata$guys <- gender=="m" #Creates a logical vector for males.

mydata$gals <- gender=="f" #Creates another for females.

print(mydata)

 

# Applies basic math to the group selected by the logic.

 

mydata$score1[gals]<- 2*q1[gals] + q2[gals]

mydata$score2[gals]<- 3*q1[gals] + q2[gals]

mydata$score1[guys]<-20*q1[guys] + q2[guys]

mydata$score2[guys]<-30*q1[guys] + q2[guys]

print(mydata)

 

# This step uses NULL to delete the guys & gals vars.

mydata$gals <- mydata$guys <- NULL

print(mydata)

 

This is how I modified the code from above to try to do what I wanted:

 

StandFTI <- cbind(StandFTI, score1 = NA, score2 = NA)

 attach(StandFTI)

StandFTI$Stand <- FTIStandKey=="NAH6005-001" #Creates a logical
vector for
males.

StandFTI$score1[Stand]<- "TEST"

print(StandFTI)

 

But I get the warning message: 

 

In `[<-.factor`(`*tmp*`, guys, value = "TEST") :

  invalid factor level, NAs generated

 

Any help will be greatly appreciated!

 

Thanks,

 

Randy Cass

 

 


	[[alternative HTML version deleted]]

> What I want to do is to be able to select certain FTIStandKey's and
change
> values in the dataset.  For example, I would like to select by
> FTIStandKey=="NAH6253-003", and change the entries for
CoverType="
> XSOP-9999-C " and change the V_SpGrp=="T", and also change
V_Origin=="T".  I
> only want to change the record for the one FTIStandKey.  All the other
> records need to remain the exact same.  I have spent a good bit of time
> trying to figure this out but with no success.   Any help would be greatly
> appreciated.
Do not know if the following will work, and it is not tested since you 
don't give a reproducible example, see ?dput and the Posting Guide for 
this list. There are several ways of doing what you want, some of which 
are  possibly more transparent than this:

mydata[mydata$FTIStandKey == "NAH6253-003",
    c("CoverType", "V_SpGrp", "V_Origin")] <-
        list(" XSOP-9999-C ", "T", "T")


This relies on the fact that if your variables are factors, that they 
include the replacement levels already (including the fact that you have 
spaces on either side of the string for CoverType... You might also see 
?ifelse for another method of dealing with stuff like this.