Details
-
Improvement
-
Status: Closed
-
Minor
-
Resolution: Fixed
-
None
-
None
-
None
-
None
-
Operating System: All
Platform: Other
-
35028
Description
Assume from a struts-tomcate web-app I get a large file that I want to store in
a DB. To save DB permanent storage, I want to compress this. To be portable, I
rather want to do that in java than on the DB-side with some proprietary commands.
In order not to have to load into the RAM at least the entire compressed output
by e.g. creating a ByteArrayOutputStream with for example
java.util.zip.GZIPOutputStream or java.util.zip.DeflaterOutputStream and then
converting it back to an InputStream for the PreparedStatement
stmt.setBinaryStream(int pos,
InputStream stream, int length), I envision an InputStream converter that is
memory efficient.
I started implementing it myself and got it working at least for
single-byte reads. But it is far from production-readiness - perhaps I should
rather extend java.io.PipedInputStream than trying to do my own efficient
buffer management and do all the synchronize() blocking etc.
==> shouldn't something go into org.apache.commons.compress.zip?
There are certainly other applications for such a class.
P.S.: Alternatively, there is a "OutputStream java.sql.Blob.setBinaryStream(long
pos)" but for example in MySQL that "updatable BLOB that can update in-place" is
not yet implemented. Doing proprietary SQL-"COMPRESS" as per
http://dev.mysql.com/doc/mysql/en/string-functions.html is probably an even less
preferrable option.