Daniel Mesejo

Mismatch of split_conditions loading via `json.load` and `load_model`

2

There is a mismatch between manually loading the split_conditions via `json.load` and via `model.load_model`: ```python import xgboost as xgb import json with open("model-single-tree.json") as infile: data = json.load(infile) json_trees =...

## Description of changes Implement the operation TableUnnest for DataFusion ## Issues closed N/A

## Description of changes These changes solve the issue referenced by binding the column of the original left table onto the filtered table (the one resulting from applying the tolerance...

### What happened? The following code snippet produces the wrong output: ```python import ibis con = ibis.duckdb.connect() left = con.create_table("left2", { "id": [1, 2], "statement_date": ["2024-01-31", "2024-02-28"], }).mutate(statement_date=ibis._.statement_date.cast("timestamp")) right =...

fix(duckdb): implement IGNORE NULLS for FirstValue and LastValue

2

## Description of changes Reported in #11726, this PR set the base changes for FirstValue and LastValue. Namely: 1. Adding the `include_null` field to both classes 2. Updating the function...

## Description of changes Fix precision loss in Divide for the SQLite backend ## Issues closed * Resolves #11743

## Description of changes Add a compilation rule for the Strftime op via date_format ## Issues closed N/A

## Description of changes Issue #11709 reported that the DuckDB backend was creating two tables from a Python object; the same happens to the SQLite backend. These changes resolved that...

fix: add order_by to first and last

11

## Description of changes N/A ## Issues closed * Resolves #11656

## Description of changes Change the View field child to parent as per the TODO comment ## Issues closed N/A