From 02821c780a57767b520807d8227950474cb4bb81 Mon Sep 17 00:00:00 2001
From: Charles Turner <charles.turner@anu.edu.au>
Date: Thu, 14 Nov 2024 13:00:39 +0800
Subject: [PATCH] Updated DfFileCatalog.add and related tests to describe
 errors relating to columns with iterables in a more satisfactory manner

---
 src/intake_dataframe_catalog/core.py | 29 ++++++++++++++++++++++++----
 tests/test_core.py                   | 12 +++++++++++-
 2 files changed, 36 insertions(+), 5 deletions(-)

diff --git a/src/intake_dataframe_catalog/core.py b/src/intake_dataframe_catalog/core.py
index 3e31dcf..caea211 100644
--- a/src/intake_dataframe_catalog/core.py
+++ b/src/intake_dataframe_catalog/core.py
@@ -266,6 +266,11 @@ def add(
         overwrite: bool, optional
             If True, overwrite all existing entries in the dataframe catalog with name_column entries
             that match the name of this source. Otherwise the entry is appended to the dataframe catalog.
+
+        Raises
+        ------
+        DfFileCatalogError
+            If the source cannot be added to the dataframe catalog.
         """
 
         metadata = metadata or {}
@@ -293,11 +298,27 @@ def add(
             # Check that new entries contain iterables when they should
             entry_iterable_columns = _columns_with_iterables(row)
             if entry_iterable_columns != self.columns_with_iterables:
-                raise DfFileCatalogError(
-                    f"Cannot add entry with iterable metadata columns: {entry_iterable_columns} "
-                    f"to dataframe catalog with iterable metadata columns: {self.columns_with_iterables}. "
-                    " Please ensure that metadata entries are consistent."
+                missing_iterable_cols = set(self.columns_with_iterables) - set(
+                    entry_iterable_columns
                 )
+                unexpected_iterable_cols = set(entry_iterable_columns) - set(
+                    self.columns_with_iterables
+                )
+
+                if missing_iterable_cols:
+                    err_msg = (
+                        f"Expected additional iterable metadata columns: {list(self.columns_with_iterables)}. "
+                        f"Unable to add entry with iterable metadata columns '{list(entry_iterable_columns)}' to dataframe catalog: "
+                        f"columns {list(missing_iterable_cols)} must be iterable to ensure metadata entries are consistent."
+                    )
+                elif unexpected_iterable_cols:
+                    err_msg = (
+                        f"Expected iterable metadata columns: {list(self.columns_with_iterables)}. "
+                        f"Unable to add entry with metadata columns '{list(entry_iterable_columns)}' to dataframe catalog: "
+                        f"columns {list(unexpected_iterable_cols)} must not be iterable to ensure metadata entries are consistent."
+                    )
+
+                raise DfFileCatalogError(err_msg)
 
             if set(self.columns) == set(row.columns):
                 if (
diff --git a/tests/test_core.py b/tests/test_core.py
index 1aad549..253762e 100644
--- a/tests/test_core.py
+++ b/tests/test_core.py
@@ -331,7 +331,17 @@ def test_catalog_add(catalog_path, source_path):
 
     with pytest.raises(DfFileCatalogError) as excinfo:
         cat.add(gistemp, metadata={"realm": "atmos", "variable": "tas"})
-    assert "Cannot add entry with iterable metadata columns" in str(excinfo.value)
+    assert (
+        str(excinfo.value)
+        == "Expected additional iterable metadata columns: ['variable']. Unable to add entry with iterable metadata columns '[]' to dataframe catalog: columns ['variable'] must be iterable to ensure metadata entries are consistent."
+    )
+
+    with pytest.raises(DfFileCatalogError) as excinfo:
+        cat.add(gistemp, metadata={"realm": ["atmos"], "variable": ["tas"]})
+    assert (
+        str(excinfo.value)
+        == "Expected iterable metadata columns: ['variable']. Unable to add entry with metadata columns '['realm', 'variable']' to dataframe catalog: columns ['realm'] must not be iterable to ensure metadata entries are consistent."
+    )
 
     with pytest.raises(DfFileCatalogError) as excinfo:
         cat.add(gistemp, metadata={"foo": "bar", "variable": ["tas"]})