[R] matched samples, dataframe, panel data

[arun]

Hi,
I changed the fun1().  Now, it should be possible to get all the possible combinations within each group.


final3New<-read.table(file="real_data_cecilia.txt",sep="\t",header=T)
final3New1<-read.csv("real_data_cecilia_new.csv")
fun1New<- function(dat,percent,number){
    lst1<- split(dat,list(dat$year,dat$industry))
    lst2<- lst1[lapply(lst1,nrow)>1]
    lst3<- lapply(lst2,function(x) {
                    CombN1<-combn(seq_len(nrow(x)),2)
                    lapply(split(CombN1,col(CombN1)),function(y){
                            x1<-x[y,]
                            x1[sum(x1$dummy)==1,]
                            })
                    })

        lst4<- lapply(lst3,function(x) x[lapply(x,nrow)>0])
    lst5<- lst4[lapply(lst4,length)>0]
    lst6<- lapply(lst5,function(x){
                   lapply(x,function(y){
                    x1<- abs(diff(y$dimension))< number
                    x2<- y$dimension[2]+ (y$dimension[2]*percent)
                    x3<- y$dimension[2]- (y$dimension[2]*percent)
                    x4<- (y$dimension[1] < x2) & (y$dimension[1] > x3)
                    y[x4 & x1,]
                    })
                    }
                    )
    lst7<- lapply(lst6,function(x) x[lapply(x,nrow)>0])
    lst8<- lst7[lapply(lst7,length)>0]
    res<- do.call(rbind,lapply(lst8,function(x){
                           do.call(rbind,x)
                        }))
    row.names(res)<- 1:nrow(res)
    res
    }    
                    
##Applying fun1New
res5Percent<- fun1New(final3New,0.05,50)
dim(res5Percent)
#[1] 718   5
res5PercentHigh<- fun1New(final3New,0.05,500000)
 dim(res5PercentHigh)
#[1] 2788    5

res5Percent1<- fun1New(final3New1,0.05,50)
dim(res5Percent1)
#[1] 870   5
res5Percent1High<- fun1New(final3New1,0.05,500000)
dim(res5Percent1High)
#[1] 2902    5

res10Percent<- fun1New(final3New,0.10,200)
dim(res10Percent)
#[1] 2928    5
res10Percent1<- fun1New(final3New1,0.10,200)
dim(res10Percent1)
#[1] 3092    5

fun3<- function(dat){
          indx<- duplicated(dat)
      dat1<- subset(dat[indx,],dummy==1)
      dat0<- subset(dat[indx,],dummy==0)
      indx1<- as.numeric(row.names(dat1))
     indx11<- sort(c(indx1,indx1+1))
     indx0<- as.numeric(row.names(dat0))
     indx00<- sort(c(indx0,indx0-1))
      indx10<- sort(c(indx11,indx00))
     res <- dat[-indx10,]
    res
    }




#Applying fun3()
res5F3<- fun3(res5Percent)
dim(res5F3)
#[1] 278   5

res5F3High<- fun3(res5PercentHigh)
dim(res5F3High)
#[1] 546   5

res5F3_1<- fun3(res5Percent1)
#[1] 302   5
res5F3High_1<- fun3(res5Percent1High)
dim(res5F3High_1)
#[1] 570   5

res10F3<- fun3(res10Percent)
dim(res10F3)
#[1] 462   5
res10F3_1<- fun3(res10Percent1)
#[1] 474   5
nrow(subset(res5F3,dummy==0))
#[1] 139
 nrow(subset(res5F3,dummy==1))
#[1] 139


 nrow(subset(res5F3High,dummy==1))
#[1] 273
 nrow(subset(res5F3High,dummy==0))
#[1] 273


 nrow(subset(res10F3,dummy==0))
#[1] 231
 nrow(subset(res10F3,dummy==1))
#[1] 231
 nrow(subset(res10F3_1,dummy==1))
#[1] 237
 nrow(subset(res10F3_1,dummy==0))
#[1] 237
 dim(unique(res5F3))
#[1] 278   5
dim(unique(res5F3High))
#[1] 546   5

 dim(unique(res10F3_1))
#[1] 474   5
 dim(unique(res10F3))
#[1] 462   5
A.K.



________________________________
From: Cecilia Carmo <cecilia.carmo at ua.pt>
To: arun <smartpink111 at yahoo.com> 
Sent: Friday, June 14, 2013 10:44 AM
Subject: me again




There some matchs that are missing. That is, it is possible to have more matchs.
I'm sending you a sript and the data.

Than you.
Cecília