Fixed bug where nested json inside pandas wouldn't be ingested correctly

CHANGELOG.md

@@ -7,6 +7,7 @@ and this project adheres to [Semantic Versioning](https://semver.org/spec/v2.0.0
 
 ## Unreleased
 
+### Changed
 - [BREAKING] Dropped support for python version 3.7, as it is on EOL for over year.
 - [BREAKING] Aligned the Connection String Builder keywords with the rest of the SDKs.
 This means that some keywords were removed, and they will no longer be parsed as part of the Connection String.  
@@ -17,7 +18,12 @@ The following keywords have been removed:
     - `interactive_login`
     - `az_cli`
 
+### Fixed
+- Fixed issues with nested pandas dataframes ingestion.
+
 ## [4.6.3] - 2025-01-08
+
+### Fixed
 - Explicitly export members in `__init__.py` via `__all__`
 
 ## [4.6.2] - 2025-01-07

azure-kusto-ingest/azure/kusto/ingest/base_ingest_client.py

@@ -1,22 +1,20 @@
+import gzip
 import ipaddress
 import os
 import tempfile
 import time
 import uuid
 from abc import ABCMeta, abstractmethod
-from copy import copy
 from enum import Enum
 from io import TextIOWrapper
 from typing import TYPE_CHECKING, Union, IO, AnyStr, Optional, Tuple
 from urllib.parse import urlparse
 
 from azure.kusto.data.data_format import DataFormat
 from azure.kusto.data.exceptions import KustoClosedError
-
 from .descriptors import FileDescriptor, StreamDescriptor
 from .ingestion_properties import IngestionProperties
 
-
 if TYPE_CHECKING:
     import pandas
 
@@ -101,12 +99,15 @@ def set_proxy(self, proxy_url: str):
         if self._is_closed:
             raise KustoClosedError()
 
-    def ingest_from_dataframe(self, df: "pandas.DataFrame", ingestion_properties: IngestionProperties) -> IngestionResult:
+    def ingest_from_dataframe(
+        self, df: "pandas.DataFrame", ingestion_properties: IngestionProperties, data_format: Optional[DataFormat] = None
+    ) -> IngestionResult:
         """Enqueue an ingest command from local files.
         To learn more about ingestion methods go to:
         https://docs.microsoft.com/en-us/azure/data-explorer/ingest-data-overview#ingestion-methods
         :param pandas.DataFrame df: input dataframe to ingest.
         :param azure.kusto.ingest.IngestionProperties ingestion_properties: Ingestion properties.
+        :param DataFormat data_format: Format to convert the dataframe to - Can be DataFormat.CSV, DataFormat.JSOn or None. If not specified, it will try to infer it from the mapping, if not found, it will default to JSON.
         """
 
         if self._is_closed:
@@ -117,12 +118,28 @@ def ingest_from_dataframe(self, df: "pandas.DataFrame", ingestion_properties: In
         if not isinstance(df, DataFrame):
             raise ValueError("Expected DataFrame instance, found {}".format(type(df)))
 
-        file_name = "df_{id}_{timestamp}_{uid}.csv.gz".format(id=id(df), timestamp=int(time.time()), uid=uuid.uuid4())
-        temp_file_path = os.path.join(tempfile.gettempdir(), file_name)
+        is_json = True
 
-        df.to_csv(temp_file_path, index=False, encoding="utf-8", header=False, compression="gzip")
+        # If we are given CSV mapping, or the mapping format is explicitly set to CSV, we should use CSV
+        if not data_format:
+            if ingestion_properties is not None and (ingestion_properties.ingestion_mapping_type == DataFormat.CSV):
+                is_json = False
+        elif data_format == DataFormat.CSV:
+            is_json = False
+        elif data_format == DataFormat.JSON:
+            is_json = True
+        else:
+            raise ValueError("Unsupported format: {}. Supported formats are: CSV, JSON, None".format(data_format))
 
-        ingestion_properties.format = DataFormat.CSV
+        file_name = "df_{id}_{timestamp}_{uid}.{ext}.gz".format(id=id(df), timestamp=int(time.time()), uid=uuid.uuid4(), ext="json" if is_json else "csv")
+        temp_file_path = os.path.join(tempfile.gettempdir(), file_name)
+        with gzip.open(temp_file_path, "wb") as temp_file:
+            if is_json:
+                df.to_json(temp_file, orient="records", date_format="iso", lines=True)
+                ingestion_properties.format = DataFormat.JSON
+            else:
+                df.to_csv(temp_file_path, index=False, encoding="utf-8", header=False)
+                ingestion_properties.format = DataFormat.CSV
 
         try:
             return self.ingest_from_file(temp_file_path, ingestion_properties)

azure-kusto-ingest/tests/test_e2e_ingest.py

@@ -529,15 +529,20 @@ async def test_streaming_ingest_from_dataframe(self):
             "xtextWithNulls",
             "xdynamicWithNulls",
         ]
-        rows = [
-            [0, "00000000-0000-0000-0001-020304050607", 0.0, 0.0, 0, 0, 0, 0, 0, 0, 0, 0, "2014-01-01T01:01:01Z", "Zero", "Zero", "0", "00:00:00", None, ""]
-        ]
+
+        guid = uuid.uuid4()
+
+        dynamic_value = ["[email protected]", "[email protected]", "[email protected]"]
+        rows = [[0, str(guid), 0.0, 0.0, 0, 0, 0, 0, 0, 0, 0, 0, "2014-01-01T01:01:01Z", "Zero", "Zero", "0", "00:00:00", None, dynamic_value]]
         df = DataFrame(data=rows, columns=fields)
-        ingestion_properties = IngestionProperties(database=self.test_db, table=self.test_table, flush_immediately=True, data_format=DataFormat.CSV)
+        ingestion_properties = IngestionProperties(database=self.test_db, table=self.test_table, flush_immediately=True)
         self.ingest_client.ingest_from_dataframe(df, ingestion_properties)
 
         await self.assert_rows_added(1, timeout=120)
 
+        a = self.client.execute(self.test_db, f"{self.test_table} | where rowguid == '{guid}'")
+        assert a.primary_results[0].rows[0]["xdynamicWithNulls"] == dynamic_value
+
     @pytest.mark.asyncio
     async def test_streaming_ingest_from_blob(self, is_managed_streaming):
         ingestion_properties = IngestionProperties(

azure-kusto-ingest/tests/test_kusto_ingest_client.py

@@ -507,7 +507,7 @@ def test_simple_ingest_from_dataframe(self, mock_pid, mock_time, mock_uuid, mock
         )
 
         ingest_client = ingest_client_class("https://ingest-somecluster.kusto.windows.net")
-        ingestion_properties = IngestionProperties(database="database", table="table", data_format=DataFormat.CSV)
+        ingestion_properties = IngestionProperties(database="database", table="table")
 
         from pandas import DataFrame
 
@@ -518,11 +518,11 @@ def test_simple_ingest_from_dataframe(self, mock_pid, mock_time, mock_uuid, mock
         result = ingest_client.ingest_from_dataframe(df, ingestion_properties=ingestion_properties)
         assert result.status == IngestionStatus.QUEUED
 
-        expected_url = "https://storageaccount.blob.core.windows.net/tempstorage/database__table__11111111-1111-1111-1111-111111111111__df_{0}_100_11111111-1111-1111-1111-111111111111.csv.gz?".format(
+        expected_url = "https://storageaccount.blob.core.windows.net/tempstorage/database__table__11111111-1111-1111-1111-111111111111__df_{0}_100_11111111-1111-1111-1111-111111111111.json.gz?".format(
             id(df)
         )
 
-        assert_queued_upload(mock_put_message_in_queue, mock_upload_blob_from_stream, expected_url)
+        assert_queued_upload(mock_put_message_in_queue, mock_upload_blob_from_stream, expected_url, format="json")
 
     @responses.activate
     @patch("azure.kusto.ingest.managed_streaming_ingest_client.ManagedStreamingIngestClient.MAX_STREAMING_SIZE_IN_BYTES", new=0)