[datatable-help] Basic join question

Yang Zhang yanghatespam at gmail.com
Thu Aug 18 03:27:16 CEST 2011


The docs say:

"When i is a data.table, x must have a key. i is joined to x using the
key and the rows in x that match are returned. An equi-join is
performed between each column in i to each column in x's key. The
match is a binary search in compiled C in O(log n) time. If i has less
columns than x's key then many rows of x may match to each row of i.
If i has more columns than x's key, the columns of i not involved in
the join are included in the result. If i also has a key, it is i's
key columns that are used to match to x's key columns and a binary
merge of the two tables is carried out."

Some additional quick questions:

1. *Which* columns of i are used in the join (assuming no keys are
set)? Is it just left-to-right?

2. When there are two columns with the same name in x and i (which
aren't being used as join keys), is just the one from x kept?


On Wed, Aug 17, 2011 at 6:26 PM, Yang Zhang <yanghatespam at gmail.com> wrote:
> I'm going to continue here since the question is a bit more
> complicated and SO isn't the best forum for back-and-forth.
>
> If I'm trying to do a join where I'm trying to aggregate counts
> (including 0s for nomatches), is there something more concise than the
> following, which is what I'm currently using since it works?
>
> # assume dt is a data.frame(user_id=..., age=...)
> y = dt[, list(count=length(age)), by=user_id]
> key(y) = 'user_id'
> y = y[J(unique(x$user_id))]
> y$count[is.na(y$count)] = 0
>
> I tried:
>
>> key(y) = 'user_id'
>> y = y[J(unique(x$user_id)), list(count=length(age))]
>> summary(y$count)
>     Min.   1st Qu.    Median      Mean   3rd Qu.      Max.
>     1.00      1.00      1.00     75.55      5.00 127200.00
>> dim(y)
> [1] 7655    2
>
> which gives me the right number of output rows but none of the lengths
> are 0, presumably because length(NA) == 1. (There are definitely users
> in x that are not in y.)
>
> But then when I tried (and there are no NAs in y$age):
>
>> count = function(x) if (any(is.na(x))) integer(0) else length(x)
>> key(y) = 'user_id'
>> y = y[J(unique(x$user_id)), list(count=count(age))]
>> summary(y$count)
>    Min.  1st Qu.   Median     Mean  3rd Qu.     Max.
>     1.0      2.0      6.0    160.4     21.0 127200.0
>> dim(y)
> [1] 3581    2
>
> Rows seem to be disappearing, and still the min is 1.
>
> At this point I'm pretty disoriented. Any explanation? Thanks in advance.
>
>
> On Wed, Aug 17, 2011 at 12:34 PM, Yang Zhang <yanghatespam at gmail.com> wrote:
>> Thanks, edited the question.
>>
>> On Wed, Aug 17, 2011 at 3:53 AM, Matthew Dowle <mdowle at mdowle.plus.com> wrote:
>>> Yang,
>>> Since you also asked on SO, suggest we answer there (after your edit please)
>>> :
>>> http://stackoverflow.com/questions/7090621/how-to-do-a-basic-left-outer-join-with-data-table-in-r
>>> Matthew
>>>
>>>
>>> "Yang Zhang" <yanghatespam at gmail.com> wrote in message
>>> news:CAKxBDU_o3i_+xujsCa0CmukDizctx6fnzbidOYZW9Co9w9iTvw at mail.gmail.com...
>>>> How do I do the equivalent to the following?
>>>>
>>>> with dt as (select 1 as a, 0 as b union select 1, 0 union select 2, 0
>>>> union select 2, 1 union select 3, 1 union select 3, 1),
>>>>  above as (select a, b from dt where b > .5),
>>>>  below as (select a, b from dt where b < .5)
>>>> select above.a, count(below.a) from above left outer join below on
>>>> (above.a = below.a) group by above.a;
>>>> a | count
>>>> ---+-------
>>>> 3 |     0
>>>> 2 |     1
>>>> (2 rows)
>>>>
>>>> How do I accomplish the same thing with data.tables?  This is what I
>>>> have so far:
>>>>
>>>> DT = data.table(a=as.integer(c(1,1,2,2,3,3)), b=c(0,0,0,1,1,1))
>>>> above = DT[DT$b > .5]
>>>> below = DT[DT$b < .5, list(a=a)]
>>>> key(below) = 'a'
>>>> below[above, list(count=length(a)), by=a]
>>>>
>>>> but this gives me:
>>>>
>>>>      a count
>>>> [1,]  2 1
>>>> [2,] NA 1
>>>>
>>>> Thanks in advance for any tips.
>>>>
>>>> --
>>>> Yang Zhang
>>>> http://yz.mit.edu/
>>>
>>>
>>>
>>> _______________________________________________
>>> datatable-help mailing list
>>> datatable-help at lists.r-forge.r-project.org
>>> https://lists.r-forge.r-project.org/cgi-bin/mailman/listinfo/datatable-help
>>>
>>
>>
>>
>> --
>> Yang Zhang
>> http://yz.mit.edu/
>>
>
>
>
> --
> Yang Zhang
> http://yz.mit.edu/
>



-- 
Yang Zhang
http://yz.mit.edu/


More information about the datatable-help mailing list