Fixes

Author: larsevj

pyproject.toml

@@ -67,6 +67,7 @@ dependencies = [
     "xarray",
     "xtgeo >= 3.3.0",
     "fastexcel>=0.12.1",
+    "xlsxwriter>=3.2.2",
 ]
 
 [project.scripts]

src/ert/config/design_matrix.py

@@ -187,8 +187,9 @@ def read_design_matrix(
             drop_empty_cols=True,
             drop_empty_rows=True,
             raise_if_empty=False,
+            infer_schema_length=None,
             read_options={"skip_rows": 1},
-        )
+        ).with_columns(pl.col(pl.Float32, pl.Float64).fill_nan(None))
 
         if error_list := DesignMatrix._validate_design_matrix(
             design_matrix_df, param_names
@@ -266,7 +267,8 @@ def _validate_design_matrix(
         empties = [
             f"Row {i}, column {j}"
             for i, j in zip(
-                *np.where(design_matrix.select(pl.all().is_null())), strict=False
+                *np.where(design_matrix.select(pl.all().is_null())),
+                strict=False,
             )
         ]
         if len(empties) > 0:

src/ert/enkf_main.py

@@ -167,12 +167,12 @@ def save_design_matrix_to_ensemble(
     if not set(design_matrix_df.get_column("REAL")).issubset(set(active_realizations)):
         raise KeyError("Active realizations not found in design matrix data frame.")
     for row in design_matrix_df.filter(
-        pl.col("REAL").is_in(active_realizations)
+        pl.col("REAL").is_in(pl.Series(active_realizations))
     ).to_numpy():
         ds = xr.Dataset(
             {
-                "values": ("names", row[1:]),
-                "transformed_values": ("names", row[1:]),
+                "values": ("names", list(row[1:])),
+                "transformed_values": ("names", list(row[1:])),
                 "names": param_names,
             }
         )

tests/ert/ui_tests/cli/analysis/test_design_matrix.py

@@ -5,8 +5,9 @@
 from textwrap import dedent
 
 import numpy as np
-import pandas as pd
+import polars as pl
 import pytest
+from xlsxwriter import Workbook
 
 from ert.cli.main import ErtCliError
 from ert.config import ErtConfig
@@ -16,11 +17,11 @@
 
 
 def _create_design_matrix(filename, design_sheet_df, default_sheet_df=None):
-    with pd.ExcelWriter(filename) as xl_write:
-        design_sheet_df.to_excel(xl_write, index=False, sheet_name="DesignSheet01")
+    with Workbook(filename) as wb:
+        design_sheet_df.write_excel(wb, worksheet="DesignSheet01")
         if default_sheet_df is not None:
-            default_sheet_df.to_excel(
-                xl_write, index=False, sheet_name="DefaultSheet", header=False
+            default_sheet_df.write_excel(
+                wb, worksheet="DefaultSheet", include_header=False
             )
 
 
@@ -30,14 +31,14 @@ def test_run_poly_example_with_design_matrix():
     a_values = list(range(num_realizations))
     _create_design_matrix(
         "poly_design.xlsx",
-        pd.DataFrame(
+        pl.DataFrame(
             {
                 "REAL": list(range(num_realizations)),
                 "a": a_values,
                 "category": 5 * ["cat1"] + 5 * ["cat2"],
             }
         ),
-        pd.DataFrame([["b", 1], ["c", 2]]),
+        pl.DataFrame([["b", 1], ["c", 2]], orient="row"),
     )
 
     with open("poly.ert", "w", encoding="utf-8") as fout:
@@ -132,13 +133,13 @@ def test_run_poly_example_with_design_matrix_and_genkw_merge(default_values, err
     a_values = list(range(num_realizations))
     _create_design_matrix(
         "poly_design.xlsx",
-        pd.DataFrame(
+        pl.DataFrame(
             {
                 "REAL": list(range(num_realizations)),
                 "a": a_values,
             }
         ),
-        pd.DataFrame(default_values),
+        pl.DataFrame(default_values, orient="row"),
     )
 
     with open("poly.ert", "w", encoding="utf-8") as fout:
@@ -219,23 +220,23 @@ def test_run_poly_example_with_multiple_design_matrix_instances():
     a_values = list(range(num_realizations))
     _create_design_matrix(
         "poly_design_1.xlsx",
-        pd.DataFrame(
+        pl.DataFrame(
             {
                 "REAL": list(range(num_realizations)),
                 "a": a_values,
             }
         ),
-        pd.DataFrame([["b", 1], ["c", 2]]),
+        pl.DataFrame([["b", 1], ["c", 2]], orient="row"),
     )
     _create_design_matrix(
         "poly_design_2.xlsx",
-        pd.DataFrame(
+        pl.DataFrame(
             {
                 "REAL": list(range(num_realizations)),
                 "d": num_realizations * [3],
             }
         ),
-        pd.DataFrame([["g", 4]]),
+        pl.DataFrame([["g", 4]], orient="row"),
     )
 
     with open("poly.ert", "w", encoding="utf-8") as fout: