Converting character to timestamp in dataframe

问题

I have a timestamp in a dataframe that is recognized as a character class. For some reason, I am not able to convert it to a poxis timestamp.

Here is a sample of the data.

ID    dateTime    stage
1    2016-11-01T00:00:00.000Z  4.82
2    2016-11-01T00:15:00.000Z  4.83
3    2016-11-01T00:30:00.000Z  4.84
4    2016-11-01T00:45:00.000Z  4.85
5    2016-11-01T01:00:00.000Z  4.86
6    2016-11-01T01:15:00.000Z  4.87

I have tried using the following.

format(df$dateTime, "Y%-%m-%d %h:%m")


as.Date(df$dateTime, "Y%-%m-%d %h:%m")



as.POSIXct(df$dateTime, tz="GMT")

None of the attempts above worked. After trying each, the class would always come out as character.

Any suggestions?

回答1:

We can use anytime

library(anytime)
anytime('2016-11-01T00:15:00.000Z')
#[1] "2016-11-01 00:15:00 IST"

Or use strptime

strptime('2016-11-01T00:15:00.000Z', '%Y-%m-%dT%H:%M:%OSZ')
#[1] "2016-11-01 00:15:00 IST"

回答2:

Z means it's UTC time. So one option is just to strip out the letter and convert to POSIXCT with tz="UTC" . As the dataframe is not easily reproducible, I'm just using the vector of two dates:

x<- as.POSIXct( gsub(pattern = "[A-Z]",replacement = " ",
                 x = c( "2016-11-01T00:45:00.000Z","2016-11-01T00:39:00.000Z") ) ,
                 tz="UTC" ) 
x
class(x)

来源：https://stackoverflow.com/questions/40938733/converting-character-to-timestamp-in-dataframe