[Bioc-sig-seq] Update only specific spaces of RangedData

[Patrick Aboyoun]

Christian,
I finished a round of changes to the subscript replacement operators in 
IRanges, which will be available on bioconductor.org in BioC 2.5 after 
the Wednesday builds have been pushed, and I should be able to meet your 
needs. Just to be sure, can you provide a larger context what what 
operation you are trying to perform? Some questions I have are:

1) What is the class and content of the replacement value?
2) How many columns in the RangedData values table are you looking to 
replace?
3) What determines which rows within each of the spaces you are looking 
to replace?

If you provide me with this information, I can let you know what the 
most efficient way you can perform the replacement. Some example subset 
replacement operations in place are:

values(x)[[i]][[j]] <- value # replace the jth column in the ith space 
with value
values(x)[[i]][k,j] <- value # replace the (k,j) selection in the ith 
space with value
values(x)[,j] <- value # replace columns denoted by j across all the 
spaces with value
values(x)[i,j] <- value # using LogicalList, IntegerList, or RangesList 
i, replace the selected rows within each of the spaces for columns 
denoted by j with value

What I'm not sure of at this point is how flexible I should allow the 
replacement value to be. Your feedback would be very helpful.


Cheers,
Patrick



Patrick Aboyoun wrote:
> Christian,
> Thanks for bringing this issue up. I have been meaning to improve the 
> subscript replacement functions in IRanges for some time and this 
> gives me a reason to make some improvements now. I need to put some 
> testing infrastructure in place to ensure the software is behaving as 
> expected, but the current solution I am working on for your problem 
> would look like
>
> values(x)[[i]][[j]] <- value    # set the jth column in the ith space
>
> As this operation shows, I am working on the values table of the 
> RangedData object and not the whole object. This will complement the 
> current method of selecting the jth column in the ith space
>
> values(x)[[i]][[j]]    # get the jth column in the ith space
>
> I'll let you know when this enhancement has been check-in.
>
>
> Patrick
>
>
>
> Christian Ruckert wrote:
>> I am missing a method to update only specific spaces of a RangedData 
>> object.
>>
>> Let 'x' be a 'RangedData' object.
>>
>> The two existing methods are:
>>
>> 'x[i]': Subsets 'x' by indexing into its spaces, so the result is
>>           of the same class, with a different set of spaces.
>>
>> 'x[[j]] <- value': Sets value as column 'j' in 'x', where 'j' can
>>           be a character, numeric, or logical scalar that indexes into
>>           the columns. The length of 'value' should equal 'nrow(x)'.
>>
>> I am looking for a mixture of both, something like:
>>
>> x[i][[j]] <- value
>>
>> which actually doesn't work this way.
>>
>> Any help would be appreciated,
>> Christian
>>
>> _______________________________________________
>> Bioc-sig-sequencing mailing list
>> Bioc-sig-sequencing at r-project.org
>> https://stat.ethz.ch/mailman/listinfo/bioc-sig-sequencing
>
> _______________________________________________
> Bioc-sig-sequencing mailing list
> Bioc-sig-sequencing at r-project.org
> https://stat.ethz.ch/mailman/listinfo/bioc-sig-sequencing