IndexOutOfBoundsException comes in case of empty dataset about abacusutil HOT 6 CLOSED

I agree that all the methods in DataSet should be compatible and work fine with empty DataSet.
The root cause for above issue mostly is in N.newEmptyDataSet(columnNames), not in DataSet
The implementation for newEmptyDataSet looks like:

new RowDataSet(new ArrayList<>(columnNames), new ArrayList<List<Object>>());

This is wrong. it should be:

final List<List<Object>> columnList = new ArrayList<>(columnNames.size());

for (int i = 0, size = columnNames.size(); i < size; i++) {
    columnList.add(new ArrayList<>(0));
}

return new RowDataSet(new ArrayList<String>(columnNames), columnList);

I will fix this issue in next release and add validation in DataSet to make sure sizes of column name list and column list are same.

I tested the join with empty DataSet not created by N.newEmptyDataSet. it works fine.

final DataSet dataSet1 = N.newDataSet(N.asLinkedHashMap("ACCT_ID", N.asList("1", "2"), "ACCT_NM", N.asList(11, 12)));

final DataSet dataSet2 = N.newDataSet(N.asList("ACCT_ID_1", "ACCT_NM_1"), N.asList(N.asList("id1", "nm1")));
dataSet2.println();

dataSet2.clear();

dataSet2.println();

dataSet2.updateColumn("ACCT_ID_1", Fn.<String, Long> f(Long::valueOf));

dataSet1.innerJoin(dataSet2, N.asMap("ACCT_ID", "ACCT_ID_1")).println();
dataSet1.fullJoin(dataSet2, N.asMap("ACCT_ID", "ACCT_ID_1")).println();

please let me know if you have different use case/test scenarios for this issue.

from abacusutil.

Yah, you are right, this issue is occurring in case of creating dataSet using N.newEmptyDataSet.
else things works fine. will wait for the new release :)

Happy coding 👍

can you please brief list of few useful classes/interfaces in abacus -> like i saw you used Fn that too has some good functions.

from abacusutil.

@landawn when I can get version 1.7.8 on maven ?

from abacusutil.

@infinite-sahil It's available now.

And as you may or will know, There are couple of alternatives for (Try.Function<String, Object, RuntimeException>) Long::valueOf, :

• cast by com.landawn.abacus.util.function.Function: (Function<String, Object>) Long::valueOf

• use: (String val) -> Long.valueOf(val)

• Fn.<String, Long> f(Long::valueOf)

• use or add new utility method. for example: Fn.parseLong(), which just added in 1.7.8 release.

from abacusutil.

By the way, if you're interested in Functional Programming, you could have a look at the section Functional Programming in this project main page

from abacusutil.

Thank you @landawn .
I verified it and its working perfectly fine.

from abacusutil.