diagram

require(plyr)

## Loading required package: plyr

require(psych)

## Loading required package: psych

df <- read.csv("matching_data.csv", header=TRUE)
df

##        ID COPD V1 V2 Gender Age
## 1  A04271 cont  0  0      1  41
## 2  A23215 cont  0  0      1  64
## 3  A23216 cont  0  0      0  73
## 4  A38144 case  0  0      1  41
## 5  A38686 cont  0  0      0  45
## 6  B16848 case  0  0      1  64
## 7  C12345 case  0  0      1  45
## 8  D12345 cont  0  0      1  41
## 9  E12345 case  0  0      1  41
## 10 F12345 case  0  0      0  41
## 11 G12345 cont  0  0      0  41
## 12 H12345 case  0  0      0  41
## 13  I2345 case  0  0      0  41
## 14 G23454 cont  0  0      0  41
## 15 F12345 cont  0  0      0  41
## 16 F12345 cont  0  0      0  41
## 17 F12345 cont  0  0      0  41
## 18 F12345 cont  0  0      0  41

# Arrange the data frame according to Age, Gender, COPD

df <- arrange(df, Age, Gender, COPD)
df

##        ID COPD V1 V2 Gender Age
## 1  F12345 case  0  0      0  41
## 2  H12345 case  0  0      0  41
## 3   I2345 case  0  0      0  41
## 4  G12345 cont  0  0      0  41
## 5  G23454 cont  0  0      0  41
## 6  F12345 cont  0  0      0  41
## 7  F12345 cont  0  0      0  41
## 8  F12345 cont  0  0      0  41
## 9  F12345 cont  0  0      0  41
## 10 A38144 case  0  0      1  41
## 11 E12345 case  0  0      1  41
## 12 A04271 cont  0  0      1  41
## 13 D12345 cont  0  0      1  41
## 14 A38686 cont  0  0      0  45
## 15 C12345 case  0  0      1  45
## 16 B16848 case  0  0      1  64
## 17 A23215 cont  0  0      1  64
## 18 A23216 cont  0  0      0  73

# Create the order of the groups in data frame for subseting data

df2 <- count(df, c('Age','Gender'))   # get the number of each group combination of Age and Gender in df
df2$seq <- 1:nrow(df2)               # create an order of group

Group <- NULL
for (i in 1:nrow(df2)){
    seq <- rep(df2$seq[i], df2$freq[i])
    Group <- c(Group, seq)
}

df$Group <- Group 

df_new = NULL
for (i in unique(df$Group)){
    df_subset <- subset(df, Group==i)
    df_subset_case <- subset(df_subset, COPD=="case")
    
    if (nrow(df_subset_case)==0){
        df_subset_case <- df_subset_case
    } else {
    df_subset_case <- df_subset_case[sample(1:nrow(df_subset_case)),]       # make the matching process ramdomly
    df_subset_case$no <- seq(from=1, to = 2*nrow(df_subset_case)-1, by = 2)
    }
    
    df_subset_cont <- subset(df_subset, COPD=="cont")
    
    if (nrow(df_subset_cont)==0){
        df_subset_cont <- df_subset_cont
    } else {
        df_subset_cont <- df_subset_cont[sample(1:nrow(df_subset_cont)),]   # # make the matching process ramdomly
        df_subset_cont$no <- seq(from=2, to = 2*nrow(df_subset_cont ), by=2)
    }
    
    df_subset <- rbind(df_subset_case, df_subset_cont)
    df_subset <- arrange(df_subset, no)
    df_new <- rbind(df_new, df_subset)
}

df_new

##        ID COPD V1 V2 Gender Age Group no
## 1   I2345 case  0  0      0  41     1  1
## 2  G23454 cont  0  0      0  41     1  2
## 3  H12345 case  0  0      0  41     1  3
## 4  F12345 cont  0  0      0  41     1  4
## 5  F12345 case  0  0      0  41     1  5
## 6  G12345 cont  0  0      0  41     1  6
## 7  F12345 cont  0  0      0  41     1  8
## 8  F12345 cont  0  0      0  41     1 10
## 9  F12345 cont  0  0      0  41     1 12
## 10 A38144 case  0  0      1  41     2  1
## 11 D12345 cont  0  0      1  41     2  2
## 12 E12345 case  0  0      1  41     2  3
## 13 A04271 cont  0  0      1  41     2  4
## 14 A38686 cont  0  0      0  45     3  2
## 15 C12345 case  0  0      1  45     4  1
## 16 B16848 case  0  0      1  64     5  1
## 17 A23215 cont  0  0      1  64     5  2
## 18 A23216 cont  0  0      0  73     6  2