How to subset a data frame by removing all rows from columns with a given string, and value less than X?

Question

I am trying to subset a data frame in R... I would like to remove all rows where the value is >0 in all columns with the word 'Blank' in it.

Ex input) df

ID OTU1 OTU2 Blank1 Blank2 1 5 0 0 2 2 3 3 0 0 3 0 9 5 0 4 2 0 0 0

Desired output

ID OTU1 OTU2 Blank1 Blank2 2 3 3 0 0 4 2 0 0 0

I can do this individually by column with df2=subset(df, subset=!(Blank1>0 | Blank2>0))

I would like to change this so that it searches for all columns with the word 'Blank' in it, then removes rows where in those columns there is a value greater than 0.

I am trying df2=subset(df, subset=!((grepl("Blank",colnames(df)))>0)) but it does not work correctly.

Answer 1

Consider using dput next time to provide a reproducible example. Given that, this code should work (but I did not test it):

df <- df[rowSums(df[, grepl("Blank", colnames(df))]) > 0, ]

Edit: This does the exact opposite of what was asked. Here you go:

df[rowSums(df[, grepl("Blank", colnames(df))]) == 0, ]

Answer 2

这也应该工作：

df[!(apply(df[,c("Blank1","Blank2")] > 0,1,sum) > 0),]

Answer 3

Using the grepl function I would use the following code:

df2 <- df[apply(df[,grepl("Blank",names(df))],1,sum)==0,]

To break that up...

apply applies a function by either row or column, the 1 argument tells it to do rows. By applying sum I'll get something non-zero if there's any zeros. If there's a possibility of negative values change sum to function(x){sum(abs(x))} which will instead take the absolute value of the cell before summing them.

Once I've applied the sum function I just check to grab only those values which are 0 :)

We wrap all that into the row argument for 'df' and we get returned only those rows that we want.

Good luck!