Hi,
Can anyone suggest how to improve the performance of this Java Method?
I'm using it in query builder to get the value of workflow variables from WorkItem.blob$context.
Originally I used two separate methods which were nested to first deserialize then extract the variable with another custom method.
After I combined the two methods into one as suggested here the report is faster, but I feel it could still be improved further.
package ext.tools;
import wt.util.WTException;
import java.io.ObjectInputStream;
import java.io.IOException;
import wt.pom.DatastoreException;
import java.io.InputStream;
import java.sql.Blob;
import wt.method.RemoteMethodServer;
import wt.pds.PDSEncoder;
public class BLOBConversionHelper {
//pass in a BLOB (Object o) and Workflow Variable name (String s)
//returns the value of the workflow variable (int)
public static int getValueFromBLOB(final Object o, final String s) throws WTException {
// deserizlize method similar to wt.pds.ConversionHelper
Object object = null;
if (RemoteMethodServer.ServerFlag) {
ObjectInputStream objectInputStream = null;
PDSEncoder encoder = null;
try {
encoder = PDSEncoder.getEncoder();
if (o instanceof String) {
objectInputStream = encoder.setInput((String)o);
}
else {
InputStream binaryStream = null;
if (o instanceof Blob) {
binaryStream = ((Blob)o).getBinaryStream();
}
else if (o instanceof InputStream) {
binaryStream = (InputStream)o;
}
if (binaryStream != null) {
objectInputStream = encoder.setInput(binaryStream);
}
}
if (objectInputStream != null) {
object = objectInputStream.readObject();
}
}
catch (Exception ex) {
throw new DatastoreException(ex);
}
finally {
if (objectInputStream != null) {
try {
objectInputStream.close();
}
catch (IOException ex2) {}
}
PDSEncoder.freeEncoder(encoder);
}
}
//find the comma after the variable name and the next comma after
//the split the string
//convert the string to an int
int val;
int indexOfs = object.toString().indexOf(s);
int indexOfcomma = object.toString().indexOf(",", indexOfs);
int indexOfEndcomma = object.toString().indexOf(",", (indexOfcomma + 1));
String checkNull = object.toString().substring((indexOfcomma + 2),(indexOfcomma + 6));
if (checkNull.equals("null")){
val = 0;
} else {
try {
val = Integer.valueOf(object.toString().substring((indexOfcomma + 2),indexOfEndcomma));
} catch (NumberFormatException ex) {
ex.printStackTrace();
val = 99999;
}
}
return val;
}
}
Example of the deserialzed BLOB before extracting the variable:
The code searches the de-serialised blob for a string like the variable name, then the next commas, then gets the substring using the position of the comma either side either side of the value.
Basically I think what I'm asking is how do I address these [(),,]
Not sure exactly for format you want the WorkItem attributes display to look like but here goes.
The image below show a report with a column for each attribute and a third column that has all attributes from the WorkItem in a single cell.
Take your pick. Both styles use Java Method and only about 10 lines of code. Report runs pretty quickly.
