Clever Geek Handbook

R: Converting from character to POSIXct loses hours and minutes

Good morning,

I am trying to convert from character to POSIXct, but when I do, I lose hours and minutes from the data.

hourlyData (dataframe)

    Login   Expo    EquityUSD   Period                    UnrealizedProfitUSD
1   252957  0.00    7.187185    2014-02-03 00:00:00.000     0.00000
2   252957  0.00    7.187772    2014-02-03 01:00:00.000     0.00000
3   252957  0.00    7.188198    2014-02-03 02:00:00.000     0.00000
4   252957  0.00    7.187825    2014-02-03 03:00:00.000     0.00000
5   252957  0.00    7.187079    2014-02-03 04:00:00.000     0.00000
6   252957  0.00    7.187079    2014-02-03 05:00:00.000     0.00000
7   252957  0.00    7.188731    2014-02-03 06:00:00.000     0.00000
8   252957  0.00    7.186279    2014-02-03 07:00:00.000     0.00000
9   252957  0.00    7.187185    2014-02-03 08:00:00.000     0.00000

when i type the class (hourlyData $ Period) i get Character

. When I try to convert this column as such hourlyData$Period = as.POSIXct(hourlyData$Period)

, however, I get the following output:
hourlyData p>

    Login   Expo    EquityUSD   Period       UnrealizedProfitUSD
1   252957  0.00    7.187185    2014-02-03   0.00000
2   252957  0.00    7.187772    2014-02-03   0.00000
3   252957  0.00    7.188198    2014-02-03   0.00000
4   252957  0.00    7.187825    2014-02-03   0.00000
5   252957  0.00    7.187079    2014-02-03   0.00000
6   252957  0.00    7.187079    2014-02-03   0.00000
7   252957  0.00    7.188731    2014-02-03   0.00000
8   252957  0.00    7.186279    2014-02-03   0.00000
9   252957  0.00    7.187185    2014-02-03   0.00000

If hours and minutes have been removed from the Period column. Does anyone know why this is happening, or how to prevent it?

thank

Mike

+3


source to share


3 answers


I would use strptime

:



char_data = c('2014-02-03 00:00:00.000', '2014-02-03 01:00:00.000', '2014-02-03 02:00:00.000')
strptime(char_data, '%Y-%m-%d %H:%M:%S')
[1] "2014-02-03 00:00:00 CET" "2014-02-03 01:00:00 CET"
[3] "2014-02-03 02:00:00 CET"
0


source


"2014-03-31 03:00:00" is not a valid date where you live, since "2014-03-31 02:59:59" is immediately followed by "2014-03-31 04: 00:00" due to daylight saving time.



0


source


The other answers hint at the problem, but don't really address it. as.POSIXct(...)

has strange behavior when a character vector is passed in with invalid times: instead of returning NA for those elements with invalid times, it as.POSIXct(...)

removes the time portion for all elements .

You can "fix" this by explicitly specifying the format specification, even if you are using the standard specification (see the last line below).

x <- sprintf('%02d:00:00', 20:25)          # 25:00:00 is not a valid time spec.
y <- sprintf('%s %s', '2018-01-01',x)      # last element has invalid time
as.POSIXct(head(y,-1))                     # works fine
## [1] "2018-01-01 20:00:00 HST" "2018-01-01 21:00:00 HST" "2018-01-01 22:00:00 HST" "2018-01-01 23:00:00 HST" "2018-01-02 00:00:00 HST"
as.POSIXct(y)                              # fails miserably
## [1] "2018-01-01 HST" "2018-01-01 HST" "2018-01-01 HST" "2018-01-01 HST" "2018-01-01 HST" "2018-01-01 HST"
as.POSIXct(y, tz='UTC')                    # tz does not fix this...
## [1] "2018-01-01 UTC" "2018-01-01 UTC" "2018-01-01 UTC" "2018-01-01 UTC" "2018-01-01 UTC" "2018-01-01 UTC"
as.POSIXct(y, format='%Y-%m-%d %H:%M:%S')  # but this does...
## [1] "2018-01-01 20:00:00 HST" "2018-01-01 21:00:00 HST" "2018-01-01 22:00:00 HST" "2018-01-01 23:00:00 HST" "2018-01-02 00:00:00 HST" NA

Running R 3.4.0 on Win 7 x64.

0


source






More articles:


All Articles