R: manipulate list of dataframes based on condition - r

Question

I consider this question difficult, it is way over my level, and I would like some help to learn how to do this myself in the future. If I'm not providing enough information, or providing unclear information, please let me know.
I have a list of dataframes:
d1<-data.frame( Data0 = c("N,R,15,P,D", "_KEY_VALUE_1", -1,1,2,3,4,5,6,7,8,9,10,11,12,13,14,15,16,17,18,19,20,21,22,23,24,25),
Data1 = c("N,15,C,D", "Garden",0.9759,0.7121,0.7376,0.7647,0.7927,0.8209,0.8487,0.8759,0.9021,0.9274,0.9518,
1,1.0249,1.0514,1.0805,1.1132,1.1508,1.1946,1.2462,1.3071,1.3793,1.4649,1.5661,1.6854,1.8254,1.9887))
d2<-data.frame(
Data0=c("N,R,2,I,D","no_flowers",-2 , 0 , 1 , 2 , 3 , 4 , 5 , 6 , 7 , 8 , 9 ,10 ,11) ,
Data1=c("N,15,C,D","Garden",0.8891 ,0.8891,0.9051,1,0.8891,0.8891,0.7907,0.8891,0.9929,0.8891,0.8891,0.8891,0.8891)
)
d3<-data.frame(Data0=c("A,X,15,P,D","_KEY_TEXT_1","Y","N","U"),
Data1=c("N,15,C,D","Garden",1.0834,1,1))
d4<-data.frame(
Data0=c("A,X,15,P,D","_KEY_TEXT_1","Y","Y","Y","Y","Y","Y","N","N","N","N","N","N"),
Data1=c("N,R,3,I,D","house_age",16,18,19,20,21,50,16,18,19,20,21,50),
Data2=c("N,15,C,D","Garden",2.2291,2.0743,1.9369,1.8148,1.7064,1.6102,2.2291,2.0743,1.9369,1.8148,1.7064,1.6102)
)
dfl<-list(d1,d2,d3,d4)
names(dfl)<-c("no_animals","no_flowers","radiation","summer_x_house_age")
If you see the first value of the first columns in each dataframe, the second letter (after the first comma) is either R or X. R stands for Ranged and X stands for not Ranged. I would like, if the letter is "R" (Ranged), to manipulate the column into two columns, i.e. I would like the result for the d1 dataframe to look like this:
For the d4 dataframe, an interaction between "summer" (Y/N) and "house age", we see that only the second column (house age) is ranged, so I would like to do the same as for d1, but for both summer=Y and summer=N.
A little bit of background on the data frames, if it makes things easier to understand:
This is the results of a glm-model I have made outside of R, and I wish to import it to R. The last column of the dataframe is always the beta-values of the regression, and the column(s) before are the variables, which sometimes are categorical (X) and sometimes continous (R). When they are continous/ranged, I must manipulate the column to get "from" and "to", because I want to use this list to calculate probabilities for some data where I have values of the regressors I have used in my glm-model. The upmost number means "from & not including infinity, to & including upmost number", second upmost number means "from & not including upmost number, to & including second upmost number", and so on.

Answer 1

Thnk I've got it.
Define a new function which looks for the key letter (R or X) and returns either a new data frame (if R) or the same data frame (if X).
Rcheck <- function(df){
# Isolate the letter being tested for R or X
key_letter <- substr(as.character(df[1,1]),3,3)
if( key_letter == "R"){ # Proceed if letter is R
# Assign new dataframe
df_new <- df
# Add new column.
df_new[,'Data0_'] <- as.character(df_new[,'Data0'])
# Shift down and add -9999 value
rows <- nrow(df_new)
df_new[,'Data0_'][4:rows] <- as.character(df_new[,'Data0'][3:(rows-1)])
df_new[,'Data0_'][3] <- "-9999"
# Take new column from the end and put it beside Data0
column1_name <- colnames(df_new)[1]
new_column_name <- colnames(df_new)[ncol(df_new)]
other_column_names <- colnames(df_new)[2:(ncol(df_new)-1)]
df_new <- df_new[,c(column1_name, new_column_name, other_column_names)]
df_new
} else{ # If letter is not R
df
}
}
Then apply this function to your list of data frames using lapply.
new_list <- lapply(dfl, Rcheck)