| 358 | |
| 359 | |
| 360 | def test_split_out_automatically(): |
| 361 | pdf = pd.DataFrame({"a": [1, 2, 3] * 1_000, "b": 1, "c": 1, "d": 1}) |
| 362 | df = from_pandas(pdf, npartitions=500) |
| 363 | q = df.groupby("a").sum() |
| 364 | assert q.optimize().npartitions == 34 |
| 365 | expected = pdf.groupby("a").sum() |
| 366 | assert_eq(q, expected) |
| 367 | |
| 368 | q = df.groupby(["a", "b"]).sum() |
| 369 | assert q.optimize().npartitions == 50 |
| 370 | expected = pdf.groupby(["a", "b"]).sum() |
| 371 | assert_eq(q, expected) |
| 372 | |
| 373 | q = df.groupby(["a", "b", "c"]).sum() |
| 374 | assert q.optimize().npartitions == 100 |
| 375 | expected = pdf.groupby(["a", "b", "c"]).sum() |
| 376 | assert_eq(q, expected) |
| 377 | |
| 378 | |
| 379 | def test_split_out_sort_values_compute(pdf, df): |