Fix cats global

.github/workflows/main.yaml

@@ -23,8 +23,13 @@ jobs:
         with:
           fetch-depth: 0
 
+      - name: Fetch upstream tags
+        run: |
+          git remote add upstream https://github.com/dask/fastparquet.git
+          git fetch upstream --tags
+
       - name: Setup conda
-        uses: mamba-org/provision-with-micromamba@main
+        uses: conda-incubator/setup-miniconda@v3
         with:
           environment-file: ci/environment-${{ matrix.CONDA_ENV }}.yml
 
@@ -53,8 +58,13 @@ jobs:
         with:
           fetch-depth: 0
 
+      - name: Fetch upstream tags
+        run: |
+          git remote add upstream https://github.com/dask/fastparquet.git
+          git fetch upstream --tags
+
       - name: Setup conda
-        uses: mamba-org/provision-with-micromamba@main
+        uses: conda-incubator/setup-miniconda@v3
         with:
           environment-file: ci/environment-${{ matrix.CONDA_ENV }}.yml
 
@@ -82,8 +92,13 @@ jobs:
         with:
           fetch-depth: 0
 
+      - name: Fetch upstream tags
+        run: |
+          git remote add upstream https://github.com/dask/fastparquet.git
+          git fetch upstream --tags
+
       - name: Setup conda
-        uses: mamba-org/provision-with-micromamba@main
+        uses: conda-incubator/setup-miniconda@v3
         with:
           environment-file: ci/environment-py310.yml
 
@@ -94,6 +109,7 @@ jobs:
           pip install hypothesis
           pip install pytest-localserver pytest-xdist pytest-asyncio
           pip install -e . --no-deps # Install fastparquet
+          pip install versioneer # Needed for pandas build
           git clone https://github.com/pandas-dev/pandas
           cd pandas
           python setup.py build_ext -j 4
@@ -117,8 +133,13 @@ jobs:
         with:
           fetch-depth: 0
 
+      - name: Fetch upstream tags
+        run: |
+          git remote add upstream https://github.com/dask/fastparquet.git
+          git fetch upstream --tags
+
       - name: Setup conda
-        uses: mamba-org/provision-with-micromamba@main
+        uses: conda-incubator/setup-miniconda@v3
         with:
           environment-file: ci/environment-py310win.yml
 

.github/workflows/test_wheel.yaml

@@ -58,7 +58,7 @@ jobs:
           python -m pip install delvewheel cython
 
       - name: Build wheels
-        uses: joerick/cibuildwheel@v2.16.5
+        uses: joerick/cibuildwheel@v2.21.3
 
       - name: Install wheels
         shell: bash -l {0}

.github/workflows/wheel.yml

@@ -50,7 +50,7 @@ jobs:
           python -m pip install delvewheel cython
 
       - name: Build wheels
-        uses: joerick/cibuildwheel@v2.18.1
+        uses: joerick/cibuildwheel@v2.21.3
 
       - uses: actions/upload-artifact@v3
         with:
@@ -105,7 +105,7 @@ jobs:
           python -m pip install delvewheel cython
 
       - name: Build wheels
-        uses: joerick/cibuildwheel@v2.18.1
+        uses: joerick/cibuildwheel@v2.21.3
 
       - uses: actions/upload-artifact@v3
         with:
@@ -160,7 +160,7 @@ jobs:
           python -m pip install delvewheel cython
 
       - name: Build wheels
-        uses: joerick/cibuildwheel@v2.18.1
+        uses: joerick/cibuildwheel@v2.21.3
 
       - uses: actions/upload-artifact@v3
         with:
@@ -215,7 +215,7 @@ jobs:
             python -m pip install delvewheel cython
 
         - name: Build wheels
-          uses: joerick/cibuildwheel@v2.18.1
+          uses: joerick/cibuildwheel@v2.21.3
 
         - uses: actions/upload-artifact@v3
           with:
@@ -246,10 +246,10 @@ jobs:
       - name: Setup Python
         uses: actions/setup-python@v5
         with:
-          python-version: "3.12"
+          python-version: "3.13"
 
       - name: Build wheels
-        uses: joerick/cibuildwheel@v2.18.1
+        uses: joerick/cibuildwheel@v2.21.3
 
       - uses: actions/upload-artifact@v3
         with:

ci/environment-py310.yml

@@ -18,6 +18,5 @@ dependencies:
   - orjson
   - ujson
   - python-rapidjson
-  - versioneer
   - meson-python
   - pyarrow

docs/source/details.rst

@@ -194,7 +194,7 @@ split data data on the values of those columns. This is done by writing a
 directory structure with *key=value* names. Multiple partition columns can
 be chosen, leading to a multi-level directory tree.
 
-Consider the following directory tree from this `Spark example <http://Spark.apache.org/docs/latest/sql-programming-guide.html#partition-discovery>`_:
+Consider the following directory tree from this `Spark example <https://spark.apache.org/docs/latest/sql-data-sources-parquet.html#partition-discovery>`_:
 
     table/
         gender=male/
@@ -261,4 +261,4 @@ the file system implementation.
 .. raw:: html
 
     <script data-goatcounter="https://fastparquet.goatcounter.com/count"
-        async src="//gc.zgo.at/count.js"></script>
+        async src="//gc.zgo.at/count.js"></script>

fastparquet/api.py

@@ -196,6 +196,7 @@ def __init__(self, fn, verify=False, open_with=default_open, root=False,
                                      "a filesystem compatible with fsspec") from e
         self.open = open_with
         self._statistics = None
+        self.global_cats = {}
 
     def _parse_header(self, f, verify=True):
         if self.fn and self.fn.endswith("_metadata"):
@@ -260,7 +261,7 @@ def columns(self):
 
     @property
     def statistics(self):
-        if self._statistics is None:
+        if not hasattr(self, '_statistics') or self._statistics is None:
             self._statistics = statistics(self)
         return self._statistics
 
@@ -318,7 +319,8 @@ def __getitem__(self, item):
         new_pf.__setstate__(
             {"fn": self.fn, "open": self.open, "fmd": fmd,
              "pandas_nulls": self.pandas_nulls, "_base_dtype": self._base_dtype,
-             "tz": self.tz, "_columns_dtype": self._columns_dtype}
+             "tz": self.tz, "_columns_dtype": self._columns_dtype,
+             "global_cats": {}}  # fresh empty dict for the slice
         )
         new_pf._set_attrs()
         return new_pf
@@ -389,7 +391,7 @@ def read_row_group_file(self, rg, columns, categories, index=None,
             f, rg, columns, categories, self.schema, self.cats,
             selfmade=self.selfmade, index=index,
             assign=assign, scheme=self.file_scheme, partition_meta=partition_meta,
-            row_filter=row_filter
+            row_filter=row_filter, global_cats=self.global_cats
         )
         if ret:
             return df
@@ -1011,7 +1013,7 @@ def __getstate__(self):
             self.fmd.row_groups = []
         return {"fn": self.fn, "open": self.open, "fmd": self.fmd,
                 "pandas_nulls": self.pandas_nulls, "_base_dtype": self._base_dtype,
-                "tz": self.tz}
+                "tz": self.tz, "global_cats": self.global_cats}
 
     def __setstate__(self, state):
         self.__dict__.update(state)