rdatetimedata.tablefreadposixct

data.table fread and ISO8601


I'm feeling a bit daft since I just cannot see the issue ...

The NEWS file says that fread can correctly identify ISO 8601 timestamps such as 2020-07-24T10:11:12.134Z (since v1.13.0). But it isn't correct:

fread(text=c("now","2020-07-24T10:11:12.134Z"), colClasses="POSIXct", sep=",")
#           now
#        <POSc>
# 1: 2020-07-24

but if I change the T to a space, it returns the correct timestamp:

fread(text=c("now","2020-07-24 10:11:12.134Z"), colClasses="POSIXct", sep=",")
#                    now
#                 <POSc>
# 1: 2020-07-24 10:11:12

This still occurs if I use tz="" or tz="UTC". (Not surprisingly, if I omit colClasses= it does not even attempt conversion.)

What am I doing wrong to get fread's internal and much-faster POSIXct converter to work? I know how to do this post-read if needed, but I have a much larger file where the use of as.POSIXct post-read is punishing.

(Windows-11, R-4.1.2, data.table-1.14.2)

If interested, verbose=TRUE does not seem to offer much insight:

fread(text=c("now","2020-07-24T10:11:12.134Z"), colClasses="POSIXct", sep=",", verbose=TRUE)
#   OpenMP version (_OPENMP)       201511
#   omp_get_num_procs()            16
#   R_DATATABLE_NUM_PROCS_PERCENT  unset (default 50)
#   R_DATATABLE_NUM_THREADS        unset
#   R_DATATABLE_THROTTLE           unset (default 1024)
#   omp_get_thread_limit()         2147483647
#   omp_get_max_threads()          16
#   OMP_THREAD_LIMIT               unset
#   OMP_NUM_THREADS                unset
#   RestoreAfterFork               true
#   data.table is using 8 threads with throttle==1024. See ?setDTthreads.
# Input contains no \n. Taking this to be a filename to open
# [01] Check arguments
#   Using 8 threads (omp_get_max_threads()=16, nth=8)
#   NAstrings = [<<NA>>]
#   None of the NAstrings look like numbers.
#   show progress = 1
#   0/1 column will be read as integer
# [02] Opening the file
#   Opening file C:\Users\r2\AppData\Local\Temp\Rtmpao7n9S\file49384a01388a
#   File opened, size = 31 bytes.
#   Memory mapped ok
# [03] Detect and skip BOM
# [04] Arrange mmap to be \0 terminated
#   \n has been found in the input and different lines can end with different line endings (e.g. mixed \n and \r\n in one file). This is common and ideal.
# [05] Skipping initial rows if needed
#   Positioned on line 1 starting: <<now>>
# [06] Detect separator, quoting rule, and ncolumns
#   Using supplied sep ','
#   No sep and quote rule found a block of 2x2 or greater. Single column input.
#   Detected 1 columns on line 1. This line is either column names or first data row. Line starts as: <<now>>
#   Quote rule picked = 0
#   fill=false and the most number of columns found is 1
# [07] Detect column types, good nrow estimate and whether first row is column names
#   Number of sampling jump points = 1 because (29 bytes from row 1 to eof) / (2 * 29 jump0size) == 0
#   Type codes (jump 000)    : C  Quote rule 0
#   'header' determined to be true because all columns are type string and a better guess is not possible
#   All rows were sampled since file is small so we know nrow=1 exactly
# [08] Assign column names
# [09] Apply user overrides on column types
#   After 0 type and 0 drop user overrides : C
# [10] Allocate memory for the datatable
#   Allocating 1 column slots (1 - 0 dropped) with 1 rows
# [11] Read the data
#   jumps=[0..1), chunk_size=1048576, total_size=24
# Read 1 rows x 1 columns from 31 bytes file in 00:00.000 wall clock time
# [12] Finalizing the datatable
#   Type counts:
#          1 : string    'C'
# =============================
#    0.000s (  0%) Memory map 0.000GB file
#    0.000s (  0%) sep='' ncol=1 and header detection
#    0.000s (  0%) Column type detection using 1 sample rows
#    0.000s (  0%) Allocation of 1 rows x 1 cols (0.000GB) of which 1 (100%) rows used
#    0.000s (  0%) Reading 1 chunks (0 swept) of 1.000MB (each chunk 1 rows) using 1 threads
#    +    0.000s (  0%) Parse to row-major thread buffers (grown 0 times)
#    +    0.000s (  0%) Transpose
#    +    0.000s (  0%) Waiting
#    0.000s (  0%) Rereading 0 columns due to out-of-sample type exceptions
#    0.000s        Total
#           now
#        <POSc>
# 1: 2020-07-24

fread(text=c("now","2020-07-24 10:11:12.134Z"), colClasses="POSIXct", sep=",", verbose=TRUE)
#   OpenMP version (_OPENMP)       201511
#   omp_get_num_procs()            16
#   R_DATATABLE_NUM_PROCS_PERCENT  unset (default 50)
#   R_DATATABLE_NUM_THREADS        unset
#   R_DATATABLE_THROTTLE           unset (default 1024)
#   omp_get_thread_limit()         2147483647
#   omp_get_max_threads()          16
#   OMP_THREAD_LIMIT               unset
#   OMP_NUM_THREADS                unset
#   RestoreAfterFork               true
#   data.table is using 8 threads with throttle==1024. See ?setDTthreads.
# Input contains no \n. Taking this to be a filename to open
# [01] Check arguments
#   Using 8 threads (omp_get_max_threads()=16, nth=8)
#   NAstrings = [<<NA>>]
#   None of the NAstrings look like numbers.
#   show progress = 1
#   0/1 column will be read as integer
# [02] Opening the file
#   Opening file C:\Users\r2\AppData\Local\Temp\Rtmpao7n9S\file493817cf4117
#   File opened, size = 31 bytes.
#   Memory mapped ok
# [03] Detect and skip BOM
# [04] Arrange mmap to be \0 terminated
#   \n has been found in the input and different lines can end with different line endings (e.g. mixed \n and \r\n in one file). This is common and ideal.
# [05] Skipping initial rows if needed
#   Positioned on line 1 starting: <<now>>
# [06] Detect separator, quoting rule, and ncolumns
#   Using supplied sep ','
#   No sep and quote rule found a block of 2x2 or greater. Single column input.
#   Detected 1 columns on line 1. This line is either column names or first data row. Line starts as: <<now>>
#   Quote rule picked = 0
#   fill=false and the most number of columns found is 1
# [07] Detect column types, good nrow estimate and whether first row is column names
#   Number of sampling jump points = 1 because (29 bytes from row 1 to eof) / (2 * 29 jump0size) == 0
#   Type codes (jump 000)    : C  Quote rule 0
#   'header' determined to be true because all columns are type string and a better guess is not possible
#   All rows were sampled since file is small so we know nrow=1 exactly
# [08] Assign column names
# [09] Apply user overrides on column types
#   After 0 type and 0 drop user overrides : C
# [10] Allocate memory for the datatable
#   Allocating 1 column slots (1 - 0 dropped) with 1 rows
# [11] Read the data
#   jumps=[0..1), chunk_size=1048576, total_size=24
# Read 1 rows x 1 columns from 31 bytes file in 00:00.000 wall clock time
# [12] Finalizing the datatable
#   Type counts:
#          1 : string    'C'
# =============================
#    0.000s (  0%) Memory map 0.000GB file
#    0.000s (  0%) sep='' ncol=1 and header detection
#    0.000s (  0%) Column type detection using 1 sample rows
#    0.000s (  0%) Allocation of 1 rows x 1 cols (0.000GB) of which 1 (100%) rows used
#    0.000s (  0%) Reading 1 chunks (0 swept) of 1.000MB (each chunk 1 rows) using 1 threads
#    +    0.000s (  0%) Parse to row-major thread buffers (grown 0 times)
#    +    0.000s (  0%) Transpose
#    +    0.000s (  0%) Waiting
#    0.000s (  0%) Rereading 0 columns due to out-of-sample type exceptions
#    0.000s        Total
#                        now
#                     <POSc>
# 1: 2020-07-24 10:11:12.134

This behavior does not change when using file= instead of text=.


Solution

  • Not sure if this is by design or not, but the culprit is keepLeadingZeros=TRUE, an option I set for other reasons.

    withr::with_options(
      list(datatable.keepLeadingZeros=FALSE), 
      fread(text=c("now","2020-07-24T10:11:12.134Z"), sep=",")
    )
    #                        now
    #                     <POSc>
    # 1: 2020-07-24 10:11:12.134
    
    withr::with_options(
      list(datatable.keepLeadingZeros=TRUE), 
      fread(text=c("now","2020-07-24T10:11:12.134Z"), sep=",")
    )
    #                         now
    #                      <char>
    # 1: 2020-07-24T10:11:12.134Z
    

    After-the-fact, I found the dupe-issue in https://github.com/Rdatatable/data.table/issues/4869, "keepLeadingZeros interferes with date recognition".


    FYI (to others and to my future self), the way I found this is to start R --vanilla --no-init --no-save, install deta.table, and start testing:

    ### in "failing" environment:
    opts <- options()
    opts <- opts[ !sapply(opts, inherits, c("list", "function")) ]
    dput(opts) # paste into the fresh R instance as opts2
    
    ### in the "fresh "environment"
    # opts2 <- structure(...) # 'opts' from above
    opts <- options()
    opts <- opts[ !sapply(opts, inherits, c("list", "function")) ]
    str(opts[ setdiff(names(opts2), names(opts)) ])
    

    and one-by-one enabling options until auto-conversion failed.