Transform to numeric a column with “NULL” values

问题

I've imported a dataset into R where in a column which should be supposed to contain numeric values are present NULL. This make R set the column class to character or factor depending on if you are using or not the stringAsFactors argument.

To give you and idea this is the structure of the dataset.

> str(data)
'data.frame':   1016 obs. of  10 variables:
 $ Date       : Date, format: "2014-01-01" "2014-01-01" "2014-01-01" "2014-01-01" ...
 $ Name       : chr  "Chi" "Chi" "Chi" "Chi" ...
 $ Impressions: chr  "229097" "3323" "70171" "1359" ...
 $ Revenue    : num  533.78 11.62 346.16 3.36 1282.28 ...
 $ Clicks     : num  472 13 369 1 963 161 1 7 317 21 ...
 $ CTR        : chr  "0.21" "0.39" "0.53" "0.07" ...
 $ PCC        : chr  "32" "2" "18" "0" ...
 $ PCOV       : chr  "3470.52" "94.97" "2176.95" "0" ...
 $ PCROI      : chr  "6.5" "8.17" "6.29" "NULL" ...
 $ Dimension  : Factor w/ 11 levels "100x72","1200x627",..: 1 3 4 5 7 8 9 10 11 1 ...

I would like to transform the PCROI column as numeric, but containing NULLs it makes this harder. I've tried to get around the issue setting the value 0 to all observations where current value is NULL, but I got the following error message:

> data$PCROI[which(data$PCROI == "NULL"), ] <- 0
Error in data$PCROI[which(data$PCROI == "NULL"), ] <- 0 : 
  incorrect number of subscripts on matrix

My idea was to change to 0 all the NULL observations and afterwards transform all the column to numeric using the as.numeric function.

回答1:

You have a syntax error:

data$PCROI[which(data$PCROI == "NULL"), ] <- 0 # will not work
data$PCROI[which(data$PCROI == "NULL")] <- 0 # will work

by the way you can say:

data$PCROI = as.numeric(data$PCROI)

it will convert your "NULL" to NA automatically.

来源：https://stackoverflow.com/questions/21173375/transform-to-numeric-a-column-with-null-values

标签

class

character

transformation

numeric