Join logic incorrect? about dataframe-js HOT 4 CLOSED

Hey. Thanks for your issue.
Indeed it seems that something is broken with the joins... I see why, my bad.

I will take a look in few hours to patch this

from dataframe-js.

Hey @PaulMest, I have created a PR #20 containing the modifications on join methods (and a randomly created .diff method) and new tests.

Please take a look and try this (branch hotfix/1.2.2 if you want to pull it).

from dataframe-js.

Here is a sample test that you can add to your test suite to verify the underlying logic is fixed.

    const dfNew1 = new DataFrame({
        id: [1, 2, 3, 5],
        value: [1.01, 2.01, 3.01, 5.01],
    }, ['id', 'value']);
    df1.show();

    const dfNew2 = new DataFrame({
        id: [1, 2, 3, 4],
        value2: [1.02, 2.02, 3.02, 4.02],
    }, ['id', 'value2']);

    assert.deepEqual(
      dfNew1.fullJoin(dfNew2, 'id').sortBy('id').toCollection(), [
        { id: 1, value: 1.01, value2: 1.02 },
        { id: 2, value: 2.01, value2: 2.02 },
        { id: 3, value: 3.01, value2: 3.02 },
        { id: 4, value: undefined, value2: 4.02 },
        { id: 5, value: 5.01, value2: undefined },
      ], 'NEW full joined.'
    );

from dataframe-js.

Thanks again @PaulMest for this user feedback.
The hotfix was merged and can now be pulled with npm.

Stay tuned with the project and contribute (with PRs) if you want !

from dataframe-js.