I think that column can be achieved by using CodelinkSet class.  

i used

fea <- CodelinkSet(f) ## f is a data file

ids <- fea$id  ## this will give the feature ids if present otherwise NA.

As i told you before, it is used for annotation. Off-course we can annotate the date by GEOquery() in an efficient way  as explained by you earlier. But if the user have a data taken from GEO , one way of annotating the data is using these feature ids to match in the GPl file to get the annotation(Gene name and Gene symbol and so on) .  I am not sure if there are  other ways also to do annotation of data present on GEO and if the GEO data is processed by your codelink package.

if the feature ids are not use full at all, then why every data on GEO has feature id.?? they could have easily neglected it and used probe name or probe id instead.

Anyway, If i process the raw data, my interest is to get a data frame containing gene name or symbol, signal intensity, SNR and call for each individual GSM file like below

Gene Symbol     Intensity      SNR         Call

KIAA0484             1.20000      0.0231      A

KIAA0495           21.0032        0.435       A
KIAA0506             2.3433        1.232       P
KIAA0515           11.023843    2.12         P

KIAA0556             10.21          3.21         P
KIAA0586               9.21           2.00        P

After the preparation of data like this for all the GSM files ,i go for meta-analysis .

This is your code as you posted for my earlier question

eset <- getGEO("GSE4797")
eset <- eset[[1]]
head(fData(eset))
# do preprocessing, e.g. with limma matrix methods. For example:
y <- backgroundCorrect.matrix(eset, offset = 50, method = "normexp")
y <- normalizeBetweenArrays(y, method = "quantile")
exprs(eset) <- y

I can get the expression from the above code .Then what about SNR here?? call, i can decide on the basis of SNR value.   Annotation can be done  by using GEO query as already well explained. Thanks