I was going to write 'Use the source, Luke', but it seems that you have
alreday found the relevant source files. I wrote a Python baed Rdata
writer and a reader sometimes ago just using that info and I am not
away of any file spec, so I know those two files are sufficient. For
what you want to do, I think you'll have to write some fairly
substantial code to process the Rdata as just XDR stream (as my python
scripts do, using the python built-in xdrlib),
because as far as I know the API you are after is not exposed - you'll
have to - and you can - cut and paste a substantial part of saveload.c
and serialize.c for that matter, of course.
I think my python-based Rdata reader would do most of what you want
(it was written for mostly diagnostic purposes as I was 'hand-crafting'
R objects in C and saving them as Rdata then read it tell me what's
wrong with them, if any) except it dumps a sort of general human
readable ascii text format rather than csv...
My sugegstion would be to use a lanaguage you are comfortable with which
comes with an xdr library, and just do it by hand...
Cook, Ian wrote:
> Hi,
>
> I am developing a tool for converting a large data frame stored in an
> uncompressed binary (XDR) RData file to a delimited text file. The data
> frame is too large to load() and extract rows from on a typical PC. I'm
> looking to parse through the file and extract individual entries without
> loading the whole thing into memory.
>
> In terms of some C source functions, instead of doing
> RestoreToEnv(R_Unserialize(connection)) which is essentially what load()
> does, I'm looking to get the documentation I would need to build a function
> "SaveToCSV()" so that I could do SaveToCSV(R_Unserialize(connection)).
>
> Where can I get documentation on the RData file format? Does a spec document
> exist?
>
> See details below.
>
> Thanks,
> Ian
>
> Ian Cook | Advanced Micro Devices, Inc. | [EMAIL PROTECTED]
>
> -------------------------
>
> Additional details:
>
> I've browsed through the relevant source code (saveload.c, serialize.c) for
> ideas.
>
> Here's a demo of the problem I'm looking to solve:
>
> # create a sample data frame
> ds <- data.frame(row1=c(1,2,3),row2=c('a','b','c'))
> # save into an uncompressed binary R dataset
> save(ds,file="ds.rdata",compress=FALSE)
> rm(ds)
>
> # Then load() can be simulated like this:
>
> # create and open a file connection
> con <- file("ds.rdata",open="rb")
> # read the first 5 characters
> readChar(con,5)
> # unserialize the remainder and restore to the environment
> ds <- unserialize(con,NULL)[["ds"]]
> close(con)
>
> But this takes up too much memory if the data set is too big. I can read in
> the file character-by-character, i.e. using readChar(), but it's obvious that
> the file format is not trivial. readChar(con,10000) for this demo yields:
>
> [EMAIL
> PROTECTED]@\b\0\0\0\0\0\0\0\0\003\r\0\0\0\003\0\0\0\001\0\0\0\002\0\0\0\003\0\0\004\002\0\0\0\001\0\0\020\t\0\0\0\006levels\0\0\0\020\0\0\0\003\0\0\0\t\0\0\0\001a\0\0\0\t\0\0\0\001b\0\0\0\t\0\0\0\001c\0\0\004\002\0\0\0\001\0\0\020\t\0\0\0\005class\0\0\0\020\0\0\0\001\0\0\0\t\0\0\0\006factor\0\0\0þ\0\0\004\002\0\0\0\001\0\0\020\t\0\0\0\005names\0\0\0\020\0\0\0\002\0\0\0\t\0\0\0\004row1\0\0\0\t\0\0\0\004row2\0\0\004\002\0\0\0\001\0\0\020\t\0\0\0\trow.names\0\0\0\r\0\0\0\002€\0\0\0\0\0\0\003\0\0\004\002\0\0\003ÿ\0\0\0\020\0\0\0\001\0\0\0\t\0\0\0\ndata.frame\0\0\0þ\0\0\0þ
>
> This would be parse-able if I had a file spec. Thanks.
>
> Ian Cook | Advanced Micro Devices, Inc. | [EMAIL PROTECTED]
>
> ______________________________________________
> R-devel@r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-devel
______________________________________________
R-devel@r-project.org mailing list
https://stat.ethz.ch/mailman/listinfo/r-devel
