geir@origin.no

Pi JDBC batch insertion problem, data loss.

Discussion created by geir@origin.no on Apr 4, 2013
Latest reply on Apr 5, 2013 by geir@origin.no

I've got a problem when I batch insert from a java application. The problem is that I lose data when I insert a second time.

 

Zero is the most occurring value. The values spans from 0 – 1, so it’s relative small variations.

 

I turned off the compression since it's was recommended in the “pi server manual” when working with manually entered data.

 

5707.archive.png

 

The sample data had a 20 seconds gap, but I changed it to 1 min, to get a successful insertion.

 

Ex:

  • First insertion of 189 samples, successful.
    • Retrieves the same amount of samples. (in PI SQL Commander)

     

  • Tries a second time with another data set. Insertion of 254 samples, successful.
    • Retrieves only 375 samples. (in PI SQL Commander)

The dataset is a month apart in time. Where are the rest of the samples???

 

I think it have something to do with a filter, who sorts out “bad data” or something.

 

I'm using:

 

com.osisoft.jdbc.Driver 1.3.0.0251

 

PI SQL DataAccessServer using PIOLEDB: 3.3.1.2

Outcomes