[BioC] Error loading, on Bioconductor 1.8, exprSet generated with bioconductor 1.7

[Sean Davis]

James W. MacDonald wrote:
> Hi Raffaele,
> 
> rcaloger wrote:
> 
>>However, this problem rise the question: is ti reasonable to use exprSet 
>>objects as backup?
> 
> 
> There are (at least) two ways to save the results of an analysis. The 
> first is to save your workspace as a binary .RData object, and load it 
> later if you want to do more work, etc. I have no idea if this is a 
> common thing to do, but you can run into the problems you have 
> experienced if you do so. Personally, I have never used this approach, 
> and I would bet that is true of most experienced R users.

I use a combination of approaches, saving .RData (but giving it a 
specific name and NOT saving everything in the workspace but only those 
objects that are "important"), but also saving all code to reproduce the 
analysis.  Often, I need to just reproduce a figure, or some such thing, 
and I want to do it with EXACTLY the same data that I used to produce 
the first.  This isn't always possible, but often is.

> The second is to save the code you used to do the analysis. If you want 
> to re-run the analysis or do more work, all you have to do is source() 
> the code and then continue on. The obvious downside of this approach is 
> the time required to re-run the code. However, I find that the vast 
> majority of the analyses I do can be re-run in a matter of minutes. If I 
> have something that took a long time to run, I will usually write the 
> results to disk (using write.table()), and then I can use read.table() 
> to read it back in.
> 
> Using some sort of R-aware editor such as (X)emacs/ESS makes saving your 
> code much more straightforward, not to mention all the niceties such as 
> parentheses matching, highlighting, etc.
> 
> HTH,
> 
> Jim
> 
> 
> 
> 
>>Thanks for your help
>>Raffaele
> 
> 
>