Datasets not equal but no differences found?

I have to rewrite some stored procedures so I am writing a test to make sure for every input I am getting the same output.

I have something like

from com.inductiveautomation.common.util import DatasetTestUtils

def testSP():
    orig_call = system.db.createSProcCall('orginalSP')
    system.db.execSProcCall(orig_call)
    new_call = system.db.createSProceCall('newSP')
    system.db.execSProcCall(new_call)
    orig_data = orig_call.getResultSet()
    new_data = new_call.getResultSet()
    comparison = DatasetTestUtils.equals(orig_data, new_data)
    print str(comparison)
    return orig_data, new_data

# another func 
def compareDS(ds1, ds2):
    pyDs1 = set(tuple(row) for row in system.dataset.toPyDataSet(ds1))
    pyDs2 = set(tuple(row) for row in system.dataset.toPyDataSet(ds2))
    print 'in original but not new'
    print pyDs1 - pyDs2
    print 'in new but not original'
    print pyDs2 - pyDs1

orig_data, new_data = testSP()
comapreDS(orig_data, new_data)

Print ds1/ds2 shows the same number of rows and columns. str(comparison) is false but @pturmel forsaw the future before I even finished writing the post so that's explainable.

I copied compareDS from @JordanCClark and both print statements show me an empty set - it seems like it should work.

The basic equality check for objects is "sameness", not content. You'll need to use Paul's module.

4 Likes

Will Paul's module work for 7.9?

Will probably have to write your own checks to determine equality.

1 Like

Oooh, no. Missed that tag.

I suppose you can look at how Paul's module does it and replicate it yourself. Whether in a real module or some jython.

Amazed you called that after I accidentally made the post only halfway through.

3 Likes

By sameness you mean two variables referring to the same object, not sameness in terms of the dataset shape (columns/ rows/ column names).

Edit: Just tested it this seems to be the case.

I guess then my compareDS function will do the heavy lifting of comparing, it seems a very reasonable way to do it.

2 Likes