> str(t2)
Classes ‘data.table’ and 'data.frame': 15349956 obs. of 2 variables:
$ id :Class 'integer64' num [1:15349956] 4.26e-319 4.26e-319 4.26e-319 4.26e-319 4.26e-319 ...
$ category: int 706 706 706 706 706 706 706 706 706 706 ...
- attr(*, ".internal.selfref")=<externalptr>
- attr(*, "sorted")= chr "id" "category"
> dput(head(t2))
structure(list(id = structure(c(4.26111856912241e-319, 4.26111856912241e-319,
4.26111856912241e-319, 4.26111856912241e-319, 4.26111856912241e-319,
4.26111856912241e-319), class = "integer64"), category = c(706L,
706L, 706L, 706L, 706L, 706L)), .Names = c("id", "category"), sorted = c("id",
"category"), class = c("data.table", "data.frame"), row.names = c(NA,
-6L))
> head(t2)
id category
1: 86246 706
2: 86246 706
3: 86246 706
4: 86246 706
5: 86246 706
6: 86246 706
> t2[J(86246,706), nomatch=0]
Empty data.table (0 rows) of 2 cols: id,category
为什么二进制搜索没有找到我们在t2头中看到的行?
答案 0 :(得分:2)
o
bit64::integer64
现在可用于分组和加入,#5369。感谢James Sams突出UPC和Clayton Stanley 提醒:fread()
已经能够检测并阅读integer64
一段时间了。
以上OP的例子:
t2[J(as.integer64(86246),706), nomatch=0L]
# id category
# 1: 86246 706
# 2: 86246 706
# 3: 86246 706
# 4: 86246 706
# 5: 86246 706
# 6: 86246 706