阅读下面的文本时,fread()
无法检测到第 8 列和第 9 列中的缺失值。这仅适用于默认选项 integer64="integer64"
。正确设置 integer64="double"
或 "character"
可检测 NA
。请注意,该文件在 V8 和 V9 中具有三种可能的 NA——,,
; ,,
;和NA
。附加 na.strings=c("NA","N/A","",""), sep=","
作为选项无效。
使用 read.csv()
的工作方式与 fread(integer="double")
相同。
要阅读的文本(也available as a file integer64_and_NA.csv):
2012,276,,0,"S1","001",1,,724135215,1590915056,
2012,276,2,8,"S1","001",1, ,,154598,0
2012,276,2,12,"S1","001",1,NA,5118863,21819477,
2012,276,2,0,"S1","011",8,3127133583,3127133583,9003982501,0
这是 fread()
的输出:
DT <- fread(input="integer64_and_NA.csv", verbose=TRUE, integer64="integer64", na.strings=c("NA","N/A",""," "), sep=",")
Input contains no \n. Taking this to be a filename to open
Detected eol as \r\n (CRLF) in that order, the Windows standard.
Looking for supplied sep ',' on line 4 (the last non blank line in the first 'autostart') ... found ok
Found 11 columns
First row with 11 fields occurs on line 1 (either column names or first row of data)
Some fields on line 1 are not type character (or are empty). Treating as a data row and using default column names.
Count of eol after first data row: 5
Subtracted 1 for last eol and any trailing empty lines, leaving 4 data rows
Type codes: 11114412221 (first 5 rows)
Type codes: 11114412221 (after applying colClasses and integer64)
Type codes: 11114412221 (after applying drop or select (if supplied)
Allocating 11 column slots (11 - 0 NULL)
0.000s ( 0%) Memory map (rerun may be quicker)
0.000s ( 0%) sep and header detection
0.000s ( 0%) Count rows (wc -l)
0.000s ( 0%) Column type detection (first, middle and last 5 rows)
0.000s ( 0%) Allocation of 4x11 result (xMB) in RAM
0.000s ( 0%) Reading data
0.000s ( 0%) Allocation for type bumps (if any), including gc time if triggered
0.000s ( 0%) Coercing data already read in type bumps (if any)
0.000s ( 0%) Changing na.strings to NA
0.001s Total
生成的 data.table 为:
DT
V1 V2 V3 V4 V5 V6 V7 V8 V9 V10 V11
1: 2012 276 NA 0 S1 001 1 9218868437227407266 724135215 1590915056 NA
2: 2012 276 2 8 S1 001 1 9218868437227407266 9218868437227407266 154598 0
3: 2012 276 2 12 S1 001 1 9218868437227407266 5118863 21819477 NA
4: 2012 276 2 0 S1 011 8 3127133583 3127133583 9003982501 0
在不是 integer64
的列中正确检测到 NA
值。对于 V8 和 V9,fread()
标记为 integer64,而不是 NA,我们有“9218868437227407266”。
有趣的是,str()
将 V8 和 V9 的相应值返回为 NA
:
str(DT)
Classes ‘data.table’ and 'data.frame': 4 obs. of 11 variables:
$ V1 : int 2012 2012 2012 2012
$ V2 : int 276 276 276 276
$ V3 : int NA 2 2 2
$ V4 : int 0 8 12 0
$ V5 : chr "S1" "S1" "S1" "S1"
$ V6 : chr "001" "001" "001" "011"
$ V7 : int 1 1 1 8
$ V8 :Class 'integer64' num [1:4] NA NA NA 1.55e-314
$ V9 :Class 'integer64' num [1:4] 3.58e-315 NA 2.53e-317 1.55e-314
$ V10:Class 'integer64' num [1:4] 7.86e-315 7.64e-319 1.08e-316 4.45e-314
$ V11: int NA 0 NA 0
- attr(*, ".internal.selfref")=<externalptr>
...但没有其他东西将它们视为NA
:
is.na(DT$V8)
[1] FALSE FALSE FALSE FALSE
max(DT$V8)
integer64
[1] 9218868437227407266
> max(DT$V8, na.rm=TRUE)
integer64
[1] 9218868437227407266
> class(DT$V8)
[1] "integer64"
> typeof(DT$V8)
[1] "double"
这似乎不仅仅是打印/屏幕问题,data.table
将它们视为巨大的整数:
DT[, V12:=as.numeric(V8)]
Warning message:
In as.double.integer64(V8) :
integer precision lost while converting to double
> DT
V1 V2 V3 V4 V5 V6 V7 V8 V9 V10 V11 V12
1: 2012 276 NA 0 S1 001 1 9218868437227407266 724135215 1590915056 NA 9.218868e+18
2: 2012 276 2 8 S1 001 1 9218868437227407266 9218868437227407266 154598 0 9.218868e+18
3: 2012 276 2 12 S1 001 1 9218868437227407266 5118863 21819477 NA 9.218868e+18
4: 2012 276 2 0 S1 011 8 3127133583 3127133583 9003982501 0 3.127134e+09
我是否遗漏了有关 integer64
的内容,或者这是一个错误?如上所述,我可以使用 integer64="double"
来解决,但可能会丢失一些精度,如帮助文件中所述。但意外的行为是使用默认的 integer64
...
这是在运行 Revolution R 3.0.2 的 Windows 8.1 64 位计算机以及运行 kubuntu 13.10、CRAN-R 3.0.2 的虚拟机上完成的。使用 CRAN 的最新稳定 data.table(截至 2014 年 2 月 7 日的 1.8.10)和 1.8.11(修订版 1110,2014-02-04 02:43:19,从 zip 手动安装为 r-forge)进行测试Windows 上的构建已损坏),Linux 上只有稳定的 1.8.10。 bit64 已安装并加载到两台机器上。
> sessionInfo()
R version 3.0.2 (2013-09-25)
Platform: x86_64-w64-mingw32/x64 (64-bit)
locale:
[1] LC_COLLATE=English_United States.1252 LC_CTYPE=English_United States.1252 LC_MONETARY=English_United States.1252 LC_NUMERIC=C
[5] LC_TIME=English_United States.1252
attached base packages:
[1] grid stats graphics grDevices utils datasets methods base
other attached packages:
[1] bit64_0.9-3 bit_1.1-11 gdata_2.13.2 xts_0.9-7 zoo_1.7-10 nlme_3.1-113 hexbin_1.26.3 lattice_0.20-24 ggplot2_0.9.3.1
[10] plyr_1.8 reshape2_1.2.2 data.table_1.8.11 Revobase_7.0.0 RevoMods_7.0.0 RevoScaleR_7.0.0
loaded via a namespace (and not attached):
[1] codetools_0.2-8 colorspace_1.2-4 dichromat_2.0-0 digest_0.6.4 foreach_1.4.1 gtable_0.1.2 gtools_3.2.1 iterators_1.0.6
[9] labeling_0.2 MASS_7.3-29 munsell_0.4.2 proto_0.3-10 RColorBrewer_1.0-5 reshape_0.8.4 scales_0.2.3 stringr_0.6.2
[17] tools_3.0.2
最佳答案
这个错误,#488 ,现已修复为 this commit在 data.table v1.9.5
的开发版本中,如果加载了 bit64
,则值将正确分配(并显示)为 NA
。 p>
require(data.table) # v1.9.5
require(bit64)
ans = fread("test.csv")
# V1 V2 V3 V4 V5 V6 V7 V8 V9 V10 V11
# 1: 2012 276 NA 0 S1 001 1 NA 724135215 1590915056 NA
# 2: 2012 276 2 8 S1 001 1 NA NA 154598 0
# 3: 2012 276 2 12 S1 001 1 NA 5118863 21819477 NA
# 4: 2012 276 2 0 S1 011 8 3127133583 3127133583 9003982501 0
关于r - fread() 失败,integer64 列中缺少值,我们在Stack Overflow上找到一个类似的问题: https://stackoverflow.com/questions/21627741/