[R] Problems with merge
arun
smartpink111 at yahoo.com
Wed May 29 17:05:31 CEST 2013
data1$A[8]
#[1] 1.4
data2$A[15]
#[1] 1.4
data2$A[15]==data1$A[8]
#[1] FALSE
You can check these links:
http://r.789695.n4.nabble.com/Comparing-decimal-numbers-td3251437.html
http://rwiki.sciviews.org/doku.php?id=misc:r_accuracy
A.K.
>The FAQ says "Other numbers have to be rounded to (typically) 53 binary digits accuracy."
>These numbers are clearly a bit less than that. Also, I've defined
these numbers, they were not calculated through a multiplication or
division >operation. I would hate to think that R would have trouble with
such a basic operation. I must be doing something wrong.
----- Original Message -----
From: arun <smartpink111 at yahoo.com>
To: R help <r-help at r-project.org>
Cc:
Sent: Wednesday, May 29, 2013 2:33 AM
Subject: Re: Problems with merge
HI,
Possibly R FAQ: 7.31
data1New<-data1
data1New$A<- round(data1New$A,2)
data2New<- data2
data2New$A<- round(data2New$A,2)
merge(data1New,data2New,by="A")
# A B C
#1 0.0 0.9 10.0
#2 1.1 0.6 11.1
#3 1.4 0.7 11.4
#4 3.1 0.4 13.1
#5 4.4 0.8 14.4
A.K.
Hello,
Lets say we have these two data frames:
data1<-data.frame(c(1.23,1.363332,6.43209,4.230593,3.10294,5.09333,1.1,1.4,4.4,0),seq(0,by=0.1,length.out=10))
names(data1)<-c("A","B")
data2<-data.frame(seq(0,6,by=0.1),seq(10,16,by=0.1))
names(data2)<-c("A","C")
And we would like to merge only the similar values from column "A".
merge(data1,data2,by="A",all=FALSE)
we get:
A B C
1 0.0 0.9 10.0
2 1.1 0.6 11.1
3 4.4 0.8 14.4
Unfortunately, it's missing the value of 1.4 in both data.frame columns "A".
Similarly, if we instead use:
merge(data1,data2,by="A",all=TRUE)
the value of 1.4 is duplicated while 0, 1.1, and 4.4 is not.
All of the mathematical functions work as they should on the 1.4,
yet merge seems to think that the two 1.4 in the different data frames
are completely different values.
I must be going crazy, because I use the merge function all the time and I've never run into this problem before.
Thanks.
More information about the R-help
mailing list