I am new to R.
I have a table with some missing data that I would like to update from a reference table.
Sample data table:
df1=data.frame(id=c(1:5),dob=as.Date(c("1/1/2001"
,"2/2/2002",NA,NA,NA),"%m/%d/%Y"),other_data=paste0("data",seq(1:5)))
Sample lookup table:
bd_ref<-data.frame(id=c(1:100),dob=as.Date(rep("1/1/1999"),"%m/%d/%Y"))
Results should be:
id dob other_data
1 1 2001-01-01 data1
2 2 2002-02-02 data2
3 3 1999-01-01 data3
4 4 1999-01-01 data4
5 5 1999-01-01 data5
I first identified the missing data and then tried to use the lookup function from the qdapTools package based on this answer Simple lookup to insert values in an R data frame as follows:
df1[is.na(df1$dob),"dob"]<-df1[is.na(df1$dob),"id"] %l% d_ref[,c("id","dob")]
but got the error:
Error in as.Date.numeric(value) : 'origin' must be supplied
It looks like the results of df1[is.na(df1$dob),"id"] %l% d_ref[,c("id","dob")] were not dates but negative numbers
[1] -719144 -719144 -719144
Is this the correct approach in general to solve this problem? If so, any idea why the negative numbers are being returned and what i can do to fix it? If not, any suggestions for the correct approach.
You can try something like this with library dplyr. I suggest you execute each line and see what is going on with the steps.
library(dplyr)
df <- inner_join(df1, bd_ref, by = 'id')
df$dob.x <- as.Date(ifelse(!is.na(df$dob.x), df$dob.x, df$dob.y), origin = '1970-01-01')
df <- select(df, -dob.y)
names(df)[2] <- 'dob'
df
id dob other_data
1 1 2001-01-01 data1
2 2 2002-02-02 data2
3 3 1999-01-01 data3
4 4 1999-01-01 data4
5 5 1999-01-01 data5