[R] Finding unique terms

[Tóth Dénes]

On 10/12/2018 08:58 AM, Dénes Tóth wrote:
> 
> 
> On 10/12/2018 04:36 AM, Jeff Newmiller wrote:
>> You said "add up"... so you did not mean to say that? Denes computed 
>> the mean...
> 
> Nice catch, Jeff. Of course I wanted to use 'sum' instead of 'mean'.

Oh, and one more note: If you have NAs in your columns, 'sum' is rarely 
the aggregate statistic that you are after. Probably this is why my 
subconscious statistician suggested 'mean'.

> 
> 
>>
>> On October 11, 2018 3:56:23 PM PDT, roslinazairimah zakaria 
>> <roslinaump using gmail.com> wrote:
>>> Hi Denes,
>>>
>>> It works perfectly as I want!
>>>
>>> Thanks a lot.
>>>
>>> On Fri, Oct 12, 2018 at 6:29 AM Dénes Tóth <toth.denes using kogentum.hu>
>>> wrote:
>>>
>>>>
>>>>
>>>> On 10/12/2018 12:12 AM, roslinazairimah zakaria wrote:
>>>>> Dear r-users,
>>>>>
>>>>> I have this data:
>>>>>
>>>>> structure(list(STUDENT_ID = structure(c(1L, 1L, 1L, 1L, 1L, 1L,
>>>>> 2L, 2L, 2L, 2L, 2L), .Label = c("AA15285", "AA15286"), class =
>>> "factor"),
>>>>>       COURSE_CODE = structure(c(1L, 2L, 5L, 6L, 7L, 8L, 2L, 3L,
>>>>>       4L, 5L, 6L), .Label = c("BAA1113", "BAA1322", "BAA2113",
>>>>>       "BAA2513", "BAA2713", "BAA2921", "BAA4273", "BAA4513"), class
>>> =
>>>>> "factor"),
>>>>>       PO1M = c(155.7, 48.9, 83.2, NA, NA, NA, 48.05, 68.4, 41.65,
>>>>>       82.35, NA), PO1T = c(180, 70, 100, NA, NA, NA, 70, 100, 60,
>>>>>       100, NA), PO2M = c(NA, NA, NA, 37, NA, NA, NA, NA, NA, NA,
>>>>>       41), PO2T = c(NA, NA, NA, 50, NA, NA, NA, NA, NA, NA, 50),
>>>>>       X = c(NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA), X.1 = c(NA,
>>>>>       NA, NA, NA, NA, NA, NA, NA, NA, NA, NA)), .Names =
>>> c("STUDENT_ID",
>>>>> "COURSE_CODE", "PO1M", "PO1T", "PO2M", "PO2T", "X", "X.1"), class =
>>>>> "data.frame", row.names = c(NA,
>>>>> -11L))
>>>>>
>>>>> I want to combine the same Student ID and add up all the values for
>>> PO1M,
>>>>> PO1T,...,PO2T obtained by the same ID.
>>>>
>>>> dat <- structure(list(STUDENT_ID = structure(c(1L, 1L, 1L, 1L, 1L,
>>> 1L,
>>>> 2L, 2L, 2L, 2L, 2L), .Label = c("AA15285", "AA15286"), class =
>>> "factor"),
>>>>       COURSE_CODE = structure(c(1L, 2L, 5L, 6L, 7L, 8L, 2L, 3L,
>>>>       4L, 5L, 6L), .Label = c("BAA1113", "BAA1322", "BAA2113",
>>>>       "BAA2513", "BAA2713", "BAA2921", "BAA4273", "BAA4513"), class =
>>>> "factor"),
>>>>       PO1M = c(155.7, 48.9, 83.2, NA, NA, NA, 48.05, 68.4, 41.65,
>>>>       82.35, NA), PO1T = c(180, 70, 100, NA, NA, NA, 70, 100, 60,
>>>>       100, NA), PO2M = c(NA, NA, NA, 37, NA, NA, NA, NA, NA, NA,
>>>>       41), PO2T = c(NA, NA, NA, 50, NA, NA, NA, NA, NA, NA, 50),
>>>>       X = c(NA, NA, NA, NA, NA, NA, NA, NA, NA, NA, NA), X.1 = c(NA,
>>>>       NA, NA, NA, NA, NA, NA, NA, NA, NA, NA)), .Names =
>>> c("STUDENT_ID",
>>>> "COURSE_CODE", "PO1M", "PO1T", "PO2M", "PO2T", "X", "X.1"), class =
>>>> "data.frame", row.names = c(NA,
>>>> -11L))
>>>>
>>>> # I assume you would like to add up the values with na.rm = TRUE
>>>> meanFn <- function(x) mean(x, na.rm = TRUE)
>>>>
>>>> # see ?aggregate
>>>> aggregate(dat[, c("PO1M", "PO1T", "PO2M")],
>>>>             by = dat["STUDENT_ID"],
>>>>             FUN = meanFn)
>>>>
>>>> # if you have largish or large data
>>>> library(data.table)
>>>> dat2 <- as.data.table(dat)
>>>> dat2[, lapply(.SD, meanFn),
>>>>        by = STUDENT_ID,
>>>>        .SDcols = c("PO1M", "PO1T", "PO2M")]
>>>>
>>>>
>>>> Regards,
>>>> Denes
>>>>
>>>>
>>>>>
>>>>> How do I do that?
>>>>> Thank you for any help given.
>>>>>
>>>>
>>

-- 
Dr. Tóth Dénes ügyvezető
Kogentum Kft.
Tel.: 06-30-2583723
Web: www.kogentum.hu