[R] How do I 'merge' a altered subset of a data.frame back into the same data.frame

[Marc Schwartz]

On Jun 4, 2010, at 1:09 PM, dominik beck wrote:

> Hi
> 
> Step 1: I create a data.frame called iolm.
> Step 2: I create a conditional subset i_wtr.
> Step 3: In this subset I add 0.3 to all values in the IOLM_AST column.
> Step 4: Now I am looking for the best way to ‘merge’ the altered subset back
> into the original iolm data.frame
> 
> ## STEP 1
>> iolm
>     ID IOLM_AST IOLM_AXIS
> 1     1     1.15    165.33
> 2     2     1.20     79.00
> 3     3     0.40     51.66
> 4     4     0.50     57.00
> 5     5     1.77      7.70
> 6     6     0.28     99.70
> 7     7     0.48    160.00
> 8     8     0.74     84.00
> 9     9     1.63     87.00
> 10   10     0.43    150.70
> ....
> 
> ## STEP 2
> 
>> i_wtr<-subset(iolm,IOLM_AXIS > 60 & IOLM_AXIS < 120)
>> i_wtr
>     ID IOLM_AST IOLM_AXIS
> 2     2     1.20      79.0
> 6     6     0.28      99.7
> 8     8     0.74      84.0
> 9     9     1.63      87.0
> 16   16     0.93      94.0
> 20   20     1.37      91.0
> 21   21     1.19      63.0
> ...
> 
> ## STEP 3
> 
> i_wtr$IOLM_AST <- i_wtr$IOLM_AST + 0.3
>> i_wtr
>     ID IOLM_AST IOLM_AXIS
> 2     2     1.50      79.0
> 6     6     0.58      99.7
> 8     8     1.04      84.0
> 9     9     1.93      87.0
> 16   16     1.23      94.0
> 20   20     1.67      91.0
> 21   21     1.49      63.0
> ...
> 
> ## STEP 4 – result (not what I wish to get)
> 
> newiolm<-merge(i_wtr, iolm, by.x="ID", by.y="ID", all = T)
>> newiolm
>     ID IOLM_AST.x IOLM_AXIS.x IOLM_AST.y IOLM_AXIS.y
> 1     1         NA          NA       1.15      165.33
> 2     2       1.50        79.0       1.20       79.00
> 3     3         NA          NA       0.40       51.66
> 4     4         NA          NA       0.50       57.00
> 5     5         NA          NA       1.77        7.70
> 6     6       0.58        99.7       0.28       99.70
> 7     7         NA          NA       0.48      160.00
> 8     8       1.04        84.0       0.74       84.00
> 9     9       1.93        87.0       1.63       87.00
> 10   10         NA          NA       0.43      150.70
> ...
> 
> 
> ## The result I am looking to get:
> 
>     ID   IOLM_AST   IOLM_AXIS
> 1     1       1.15      165.33
> 2     2       1.50       79.00
> 3     3       0.40       51.66
> 4     4       0.50       57.00
> 5     5       1.77       7.70
> 6     6       0.58       99.70
> 7     7       0.48       160.00
> 8     8       1.04       84.00
> 9     9       1.93       87.00
> 10   10       0.43       150.70
> ...
> 
> What is the correct way to do this?
> Thanks a lot for your help.
> Dominik


Just replace the values that meet the condition with the new values, leaving the others alone:

> iolm
   ID IOLM_AST IOLM_AXIS
1   1     1.15    165.33
2   2     1.20     79.00
3   3     0.40     51.66
4   4     0.50     57.00
5   5     1.77      7.70
6   6     0.28     99.70
7   7     0.48    160.00
8   8     0.74     84.00
9   9     1.63     87.00
10 10     0.43    150.70


# See ?ifelse and ?with

> with(iolm, ifelse(IOLM_AXIS > 60 & IOLM_AXIS < 120, 
                    IOLM_AST + 0.3, 
                    IOLM_AST))
 [1] 1.15 1.50 0.40 0.50 1.77 0.58 0.48 1.04 1.93 0.43


# Replace the original column with the above result

iolm$IOLM_AST <- with(iolm, ifelse(IOLM_AXIS > 60 & IOLM_AXIS < 120, 
                                   IOLM_AST + 0.3, 
                                   IOLM_AST))

> iolm
   ID IOLM_AST IOLM_AXIS
1   1     1.15    165.33
2   2     1.50     79.00
3   3     0.40     51.66
4   4     0.50     57.00
5   5     1.77      7.70
6   6     0.58     99.70
7   7     0.48    160.00
8   8     1.04     84.00
9   9     1.93     87.00
10 10     0.43    150.70


HTH,

Marc Schwartz