[R] parts of data frames: subset vs. [-c()]

[Peter Dalgaard]

"Stefan Th. Gries" <stgries_lists at arcor.de> writes:

> Dear all
> 
> I have a problem with splitting up a data frame called ReVerb:
> 
> » str(ReVerb)
> `data.frame':   92713 obs. of  16 variables:
>  $ CHILD    : Factor w/ 7 levels "ABE","ADA","EVE",..: 1 1 1 1 1 1 1 1 1 1 ...
>  $ AGE      : Factor w/ 484 levels "1;06.00","1;06.16",..: 43 43 43 99 99 99 99 99 99 99 ...
>  $ AGE_Q    : num  2.0 2.0 2.0 2.4 2.4 ...
>  $ INTERVALS: num  2 2 2 2.25 2.25 2.25 2.25 2.25 2.25 2.25 ...
>  $ RND      : int  34368 38311 14949 20586 72516 27186 88019 10767 114448 86146 ...
>  $ SYNTAX   : Factor w/ 17 levels "Acmp","Amats",..: 15 12 8 15 7 16 7 7 16 7 ...
>  $ LEXICAL  : Factor w/ 1643 levels "$ACHE","$ACT",..: 194 803 803 294 299 803 1562 299 679 1562 ...
>  $ MORPH    : Factor w/ 337 levels "$","$ =inf","$ =prs",..: 9 20 9 39 184 231 57 67 231 39 ...
>  $ COMPLEM  : Factor w/ 1989 levels "$","$ V PR=Lp [1.2]",..: 203 547 220 203 1101 368 1834 1667 368 1834 ...
>  $ MATRIX   : Factor w/ 906 levels "$ ???","$ be PR=Aen",..: 5 5 5 308 5 856 5 5 856 308 ...
>  $ SITUATION: Factor w/ 9 levels "[imitation of Mom: you know what I said]",..: 2 2 2 2 2 2 2 2 2 2 ...
>  $ V_ANN    : int  1 1 1 4 4 4 4 3 3 3 ...
>  $ QUEST    : int  0 0 0 0 0 0 0 0 0 0 ...
>  $ EXCL     : int  0 0 0 1 1 1 1 0 0 0 ...
>  $ U_LEN    : int  3 4 5 13 13 13 13 8 8 8 ...
>  $ UTTERANCE: Factor w/ 55113 levels "","# (be)cause he wanted to .",..: 5696 39091 52180 2262 2262 2262 2262 3593 3593 3593 ...
> 
> The level causing the problem is SYNTAX:
> 
> » as.data.frame(sort(table(SYNTAX)))
>               sort(table(SYNTAX))
> Particles                     100
> PR=N1                         144
> Amats                         271
> Trans_PR=A2                   787
> Ditrans                      1181
> Intrans_PR=A1                1399
> Acmp                         2402
> Trans_PR=V2                  2433
> CPcmps                       2769
> Vpreps                       4896
> Intrans_V0                   5182
> Trans_PR=L2                  7653
> Trans_V02                    8117
> Intrans_PR=L1                8457
> Intrans_V1                   9643
> Intrans_PR=V1               14987
> Trans_V12                   22288
> 
> 
> I would like to extract all cases where SYNTAX=="Ditrans" from ReVerb, store that in a file, and then generate ReVerb again without these cases and factor levels. My problem is probably obvious from the following lines of code:
> 
> » ditrans<-which(SYNTAX=="Ditrans")
> » ReVerb1<-ReVerb[-c(ditrans),]; dim(ReVerb1)
> [1] 91532    16
> » 
> » # ok, so the 92713-91532=1181 cases where SYNTAX=="Ditrans" have been removed, but ...
> » 
> » ReVerb1<-subset(ReVerb, SYNTAX!="Ditrans"); dim(ReVerb1)
> [1] 91528    16
> » 
> » # ... so why don't I get 91532 again as the number of rows?
> » 
> Any ideas??

The SYNTAX variable is not necessarily the same. Could you retry the
first case with 

 ditrans <- which(ReVerb$SYNTAX=="Ditrans")

? 

Otherwise, try doing a setdiff() on the rownames of the two discrepant
results and see which are the four cases that differ.

-- 
   O__  ---- Peter Dalgaard             Øster Farimagsgade 5, Entr.B
  c/ /'_ --- Dept. of Biostatistics     PO Box 2099, 1014 Cph. K
 (*) \(*) -- University of Copenhagen   Denmark          Ph:  (+45) 35327918
~~~~~~~~~~ - (p.dalgaard at biostat.ku.dk)                  FAX: (+45) 35327907