#or
library(plyr)
res<-ddply(df1,.(INDX),summarize,Debut=head(Debut,1),Fin=tail(Fin,1))
res$INDX<-factor(res$INDX,levels=unique(df1$INDX))
res[order(res$INDX),-1]
# Debut Fin
#3 24/01/1995 31/12/1997
#4 02/02/1995 12/03/1995
#1 13/03/1995 30/06/1995
#2 01/01/1996 31/01/1996
A.K.
----- Original Message -----
From: arun <smartpink111 at yahoo.com>
To: Arnaud Michel <michel.arnaud at cirad.fr>
Cc: R help <r-help at r-project.org>; Rui Barradas <ruipbarradas at sapo.pt>
Sent: Wednesday, July 17, 2013 4:14 PM
Subject: Re: [R] simplify a dataframe
Hi,
You could try:
df1[,1:2]<-lapply(df1[,1:2],as.character)
df2New<- data.frame(Deb=unique(with(df1,ave(Debut,INDX,FUN=function(x) head(x,1)))),Fin=unique(with(df1,ave(Fin,INDX,FUN=function(x) tail(x,1)))))
identical(df2New,df2)
#[1] TRUE
A.K.
----- Original Message -----
From: Arnaud Michel <michel.arnaud at cirad.fr>
To: Rui Barradas <ruipbarradas at sapo.pt>; R help <r-help at r-project.org>; arun <smartpink111 at yahoo.com>
Cc:
Sent: Wednesday, July 17, 2013 4:03 PM
Subject: Re: [R] simplify a dataframe
Thank you for the question (1)
Sorry for the imprecision for the question (2) :
Suppose the date frame df
df1 <- data.frame(
Debut =c ( "24/01/1995", "01/05/1997" ,"31/12/1997", "02/02/1995"
,"28/02/1995"
,"01/03/1995", "13/03/1995", "01/01/1996", "31/01/1996") ,
Fin = c ( "30/04/1997", "30/12/1997" ,"31/12/1997", "27/02/1995",
"28/02/1995",
"12/03/1995", "30/06/1995", "30/01/1996", "31/01/1996") ,
INDX = c(6,6,6, 11,11,11, 4, 5,5) )
I would like replace df1 by df2
df2 <- data.frame(
Deb = c("24/01/1995", "02/02/1995", "13/03/1995",
"01/01/1996") ,
Fin = c("31/12/1997", "12/03/1995", "30/06/1995",
"31/01/1996") )
Explication :
The lines 1, 2 3 of df1 (who have same value of index =6) are replaced
by only one line with
value of Debut of df2 = Debut of line 1 of df1
value of Fin of df2 = Fin of line 3 of df1
The lines 4,5,6 of df1 (who have same value of index =11) are replaced
by only one line with
value of Debut of df2 = Debut of line 4 of df1
and value of fin of df2 = Fin of line 6 of df1
The line 7 of df1 (who have same value of index =4) are replaced by only
one line with
value of Debut of df2 = Debut of line 7of df1
and value of fin of df2 = Fin of line 7of df1
==> No change
The lines 8,9 of df1 (who have same value of index =5) are replaced by
only one line with
value of Debut of df2 = Debut of line 8of df1
and value of fin of df2 = Fin of line 9 of df1
df1
Debut Fin INDX
1 24/01/1995 30/04/1997 6
2 01/05/1997 30/12/1997 6
3 31/12/1997 31/12/1997 6
4 02/02/1995 27/02/1995 11
5 28/02/1995 28/02/1995 11
6 01/03/1995 12/03/1995 11
7 13/03/1995 30/06/1995 4
8 01/01/1996 30/01/1996 5
9 31/01/1996 31/01/1996 5
Deb Fin
1 24/01/1995 31/12/1997
2 02/02/1995 12/03/1995
3 13/03/1995 30/06/1995
4 01/01/1996 31/01/1996
Thank you for your helps
Michel
Le 17/07/2013 19:57, Rui Barradas a ?crit :
Hello,
As for question (1), try the following.
y2 <- cumsum(c(TRUE, diff(x1) > 0))
identical(as.integer(y1), y2) # y1 is of class "numeric"
As for question (2) I'm not understanding it.
Hope this helps,
Rui Barradas
Em 17-07-2013 18:21, Arnaud Michel escreveu:
Hi Arun
I have two questions always about the question of symplify a dataframe
I would like
1) to transform the vector x1 into the vector y1
x1 <- c(1,1,1,-1000, 1,-1000, 1,1,1,1,1,1,-1000)
y1 <- c(1,1,1,1, 2,2, 3,3,3,3,3,3,3)
2) to transform the vectors Debut and Fin by taking into account INDX
into the two vectors Deb and Fin
Debut <- c (
"24/01/1995", "01/05/1997" ,"31/12/1997", "02/02/1995" ,"28/02/1995"
,"01/03/1995",
"13/03/1995", "01/01/1996", "31/01/1996", "24/01/1995", "01/07/1995"
,"01/09/1995",
"01/07/1997", "01/01/1998", "01/08/1998", "01/01/2000",
"17/01/2000","29/02/2000")
Fin <- c (
"30/04/1997", "30/12/1997" ,"31/12/1997", "27/02/1995", "28/02/1995",
"12/03/1995",
"30/06/1995", "30/01/1996", "31/01/1996", "30/06/1995", "31/08/1995",
"30/06/1997",
"31/12/1997", "31/07/1998", "31/12/1999", "16/01/2000", "28/02/2000",
"29/02/2000")
INDX <- c(6,6,6, 11,11,11, 4, 5,5)
Deb <- c("*24/01/1995*", "*02/02/1995*", "*13/03/1995*",
"*01/01/1996*")
Fi n <- c("*31/12/1997*", "*12/03/1995*", "*30/06/1995*",
"*31/01/1996*")
Debut Fin INDX
*24/01/1995* 30/04/1997 6
01/05/1997 30/12/1997 6
31/12/1997 *31/12/1997* 6
*02/02/1995* 27/02/1995 11
28/02/1995 28/02/1995 11
01/03/1995 *12/03/1995* 11
*13/03/1995* *30/06/1995* 4
*01/01/1996* 30/01/1996 5
31/01/1996 *31/01/1996* 5
................
Thanks for your help