Merge pull request #2246 from mabel-dev/#2245

#2245
mabel-dev · Jan 14, 2025 · 9853e78 · 9853e78
2 parents 88e6e70 + f17d286
commit 9853e78
Show file tree

Hide file tree

Showing 5 changed files with 27 additions and 5 deletions.
diff --git a/opteryx/__version__.py b/opteryx/__version__.py
@@ -1,4 +1,4 @@
-__build__ = 982
+__build__ = 983
 
 # Licensed under the Apache License, Version 2.0 (the "License");
 # you may not use this file except in compliance with the License.

diff --git a/opteryx/connectors/disk_connector.py b/opteryx/connectors/disk_connector.py
@@ -225,7 +225,9 @@ def get_dataset_schema(self) -> RelationSchema:
         if self.schema:
             return self.schema
 
-        self.schema = next(self.read_dataset(just_schema=True), None)
+        for schema in self.read_dataset(just_schema=True):
+            self.schema = schema
+            break
 
         if self.schema is None:
             if os.path.isdir(self.dataset):

diff --git a/opteryx/connectors/gcp_cloudstorage_connector.py b/opteryx/connectors/gcp_cloudstorage_connector.py
@@ -275,7 +275,9 @@ def get_dataset_schema(self) -> RelationSchema:
         number_of_blobs = sum(len(b) for b in self.blob_list.values())
 
         # Read first blob for schema inference and cache it
-        self.schema = next(self.read_dataset(just_schema=True), None)
+        for schema in self.read_dataset(just_schema=True):
+            self.schema = schema
+            break
 
         if self.schema is None:
             raise DatasetNotFoundError(dataset=self.dataset)

diff --git a/opteryx/virtual_datasets/missions.py b/opteryx/virtual_datasets/missions.py
@@ -15,6 +15,8 @@
 Space Mission dataset acquired from [Kaggle](https://www.kaggle.com/datasets/agirlcoding/all-space-missions-from-1957).
 """
 
+import datetime
+
 from orso.schema import FlatColumn
 from orso.schema import RelationSchema
 from orso.types import OrsoTypes
@@ -62,4 +64,12 @@ def schema():
 
 
 def statistics() -> RelationStatistics:
-    return RelationStatistics()
+    stats = RelationStatistics()
+
+    # fmt:off
+    stats.record_count = 4630
+    stats.lower_bounds = {'Company': 'AEB', 'Location': 'Blue Origin Launch Site, West Texas, Texas, USA', 'Price': 2.5, 'Lauched_at': datetime.datetime(1957, 10, 4, 19, 28), 'Rocket': 'ASLV', 'Rocket_Status': 'Active', 'Mission': '-TJS_6.00', 'Mission_Status': 'Failure'}
+    stats.upper_bounds = {'Company': 'i-Space', 'Location': 'Xichang Satellite Launch Center, China', 'Price': 450.0, 'Lauched_at': datetime.datetime(2022, 7, 29, 13, 28), 'Rocket': 'Zoljanah', 'Rocket_Status': 'Retired', 'Mission': 'iPStar-1', 'Mission_Status': 'Success'}
+    stats.null_count = {'Company': 0, 'Location': 0, 'Price': 3380, 'Lauched_at': 127, 'Rocket': 0, 'Rocket_Status': 0, 'Mission': 0, 'Mission_Status': 0}
+    # fmt:on
+    return stats
diff --git a/opteryx/virtual_datasets/satellite_data.py b/opteryx/virtual_datasets/satellite_data.py
@@ -70,4 +70,12 @@ def schema():
 
 
 def statistics() -> RelationStatistics:
-    return RelationStatistics()
+    stats = RelationStatistics()
+
+    # fmt:off
+    stats.record_count = 177
+    stats.lower_bounds = {'id': 1, 'planetId': 3, 'name': 'Adrastea', 'gm': -0.0, 'radius': 0.3, 'density': 0.34, 'magnitude': -12.74, 'albedo': 0.04}
+    stats.upper_bounds = {'id': 177, 'planetId': 9, 'name': 'Ymir', 'gm': 9887.834, 'radius': 2631.2, 'density': 3.528, 'magnitude': 27.0, 'albedo': 1.67}
+    stats.null_count = {'id': 0, 'planetId': 0, 'name': 0, 'gm': 0, 'radius': 0, 'density': 0, 'magnitude': 0, 'albedo': 0}
+    # fmt:on
+    return stats