Hi everyone,
Has anyone written a parser in Java for either the ASCII or binary format
produced by save()? I need to parse a single large 2D array that is
structured like this:
list(
"32609_1" = c(-9549.39231289146, -9574.07159324482, ... ),
"32610_2" = c(-6369.12526971635, -6403.99620977124, ... ),
"32618_2" = c(-2138.29095689061, -2057.9229403233, ... ),
...
)
Or, given that I'm dealing with just a single array, would it be better to
roll my own I/O using write.table or write.matrix from the MASS package?
Thanks,
David
On Wed, 5 Dec 2007, David Coppit wrote:> Hi everyone, > > Has anyone written a parser in Java for either the ASCII or binary format > produced by save()? I need to parse a single large 2D array that is > structured like this: > > list( > "32609_1" = c(-9549.39231289146, -9574.07159324482, ... ), > "32610_2" = c(-6369.12526971635, -6403.99620977124, ... ), > "32618_2" = c(-2138.29095689061, -2057.9229403233, ... ), > ... > ) > > Or, given that I'm dealing with just a single array, would it be better to > roll my own I/O using write.table or write.matrix from the MASS package?It would be much easier. The save() format is far more complex than you need. However, I would use writeBin() to write a binary file and read that in in Java, avoiding the binary -> ASCII -> binary conversion. -- Brian D. Ripley, ripley at stats.ox.ac.uk Professor of Applied Statistics, http://www.stats.ox.ac.uk/~ripley/ University of Oxford, Tel: +44 1865 272861 (self) 1 South Parks Road, +44 1865 272866 (PA) Oxford OX1 3TG, UK Fax: +44 1865 272595
David Coppit wrote:> Hi everyone, > > Has anyone written a parser in Java for either the ASCII or binary format > produced by save()?You might want to consider using the hdf5 package to save the array in HDF5 format. There are HDF5 libraries for Java as well <http://hdf.ncsa.uiuc.edu/hdf-java-html/>. I have never used them, but it works quite well for transferring data between R and Python.
On 12/5/07 12:15 PM, "Prof Brian Ripley" <ripley@stats.ox.ac.uk> wrote: On Wed, 5 Dec 2007, David Coppit wrote:> Or, given that I''m dealing with just a single array, would it be better to > roll my own I/O using write.table or write.matrix from the MASS package?It would be much easier. The save() format is far more complex than you need. However, I would use writeBin() to write a binary file and read that in in Java, avoiding the binary -> ASCII -> binary conversion. Thanks for the suggestion-writeBin works quite well. For posterity, here''s what I did: On the R side: # Assumes that there are no special values in the tofList, such as NA_REAL, # R_PosInf, R_NegInf, ISNAN, R_FINITE. See the "R Data Import/Export" manual. saveListAsBinary <- function( tofList, filename ) { outConn <- file( filename, "wb" ); for (m in 1:length(tofList)) { writeBin(names(tofList)[[m]], outConn); writeBin(length(tofList[[m]]), outConn, size = 4, endian = "big"); writeBin(tofList[[m]], outConn, size = 4, endian = "big"); } close(outConn); } saveListAsBinary(myList, "outfile.RDat"); On the Java side: public static void read_R_Output(String filename, ArrayList<String> names, ArrayList<ArrayList<Float>> data) { try { DataInputStream dataInputStream = new DataInputStream( new BufferedInputStream(new FileInputStream(filename))); boolean endOfFile = false; while (!endOfFile) { try { StringBuffer sb = new StringBuffer(); byte c; while ((c = dataInputStream.readByte()) != 0) sb.append((char)c); names.add(new String(sb)); int cols = dataInputStream.readInt(); ArrayList<Float> row = new ArrayList<Float>(cols); for (int i = 0; i < cols; i++) row.add(dataInputStream.readFloat()); data.add(row); } catch (EOFException e) { endOfFile = true; } } dataInputStream.close(); } catch (Exception e) { e.printStackTrace(); } } Regards, David [[alternative HTML version deleted]]
Dear David,
You may also consider using the biocep project' tools and frameworks. they
provide an advanced bridge that allow you to exchange between R and Java
any standard R Object and any mapped S4 object.
the object extracted to Java (an RList for you data) can be serialized to
a file (saved as a java object). you can then read the serialized object
very easily from java and this is obviously much faster than parsing and
more elegant than a "proprietary" serialization/deserialization .
the easiest way to achieve this is to create the Serialized RList via the
Virtual R Workbench (Universal IDE for R).
you can run the R workbench on any kind of OS via the biocep sources:
http://www.ebi.ac.uk/microarray-srv/frontendapp/BIOCEP_README.txt
or just use this link to get the software installed on your machine if you
are using Mac OS or windows
http://www.ebi.ac.uk/microarray-srv/frontendapp/rworkbench.jnlp
create your data via the workbench R Console (you may also use the script
editor or just read your list from a file)
go to the menu "Java", choose "Save R/Java Object to local
file"
select the name you've given to your list and choose a file name.
to retrieve your list in java, all you need is to add the RJB.jar to your
class path and use this simple code:
RList l=(RList)new ObjectInputStream(new
FileInputStream("J:/list.ser")).readObject();
and that's it.
this works for much more complex R objects (ExpressionSet..)
best wishes,
Karim
>
> On 12/5/07 12:15 PM, "Prof Brian Ripley" <ripley at
stats.ox.ac.uk> wrote:
>
> On Wed, 5 Dec 2007, David Coppit wrote:
>
>> Or, given that I'm dealing with just a single array, would it be
better
>> to
>> roll my own I/O using write.table or write.matrix from the MASS
package?
>
> It would be much easier. The save() format is far more complex than you
> need. However, I would use writeBin() to write a binary file and read
> that in in Java, avoiding the binary -> ASCII -> binary conversion.
>
> Thanks for the suggestion-writeBin works quite well. For posterity,
here's
> what I did:
>
> On the R side:
>
> # Assumes that there are no special values in the tofList, such as
> NA_REAL,
> # R_PosInf, R_NegInf, ISNAN, R_FINITE. See the "R Data
Import/Export"
> manual.
> saveListAsBinary <- function( tofList, filename )
> {
> outConn <- file( filename, "wb" );
>
> for (m in 1:length(tofList)) {
> writeBin(names(tofList)[[m]], outConn);
> writeBin(length(tofList[[m]]), outConn, size = 4, endian =
"big");
> writeBin(tofList[[m]], outConn, size = 4, endian = "big");
> }
>
> close(outConn);
> }
>
> saveListAsBinary(myList, "outfile.RDat");
>
> On the Java side:
>
> public static void read_R_Output(String filename, ArrayList<String>
> names,
> ArrayList<ArrayList<Float>> data)
> {
> try {
> DataInputStream dataInputStream = new DataInputStream(
> new BufferedInputStream(new FileInputStream(filename)));
>
> boolean endOfFile = false;
>
> while (!endOfFile) {
> try {
> StringBuffer sb = new StringBuffer();
>
> byte c;
> while ((c = dataInputStream.readByte()) != 0)
> sb.append((char)c);
>
> names.add(new String(sb));
>
> int cols = dataInputStream.readInt();
>
> ArrayList<Float> row = new ArrayList<Float>(cols);
>
> for (int i = 0; i < cols; i++)
> row.add(dataInputStream.readFloat());
>
> data.add(row);
> } catch (EOFException e) {
> endOfFile = true;
> }
> }
>
> dataInputStream.close();
> } catch (Exception e) {
> e.printStackTrace();
> }
> }
>
> Regards,
> David
>
> [[alternative HTML version deleted]]
>
> ______________________________________________
> R-help at r-project.org mailing list
> https://stat.ethz.ch/mailman/listinfo/r-help
> PLEASE do read the posting guide
> http://www.R-project.org/posting-guide.html
> and provide commented, minimal, self-contained, reproducible code.
>
Reasonably Related Threads
- DO NOT REPLY [Bug 4870] New: --link-dest results in errors for some paths containing spaces
- directing print.packageInfo to a file
- [LLVMdev] Stack roots and function parameters
- How to create a list that grows automatically
- How to create a list that grows automatically