vdk-jupyter: use vdksql for SQL cells and steps (#2729)

%%vdksql is preferred over `job_input.execute_query` in a notebook as it's generally easier to write, debug has better table visualization. So we should default to it.
vmware · Oct 2, 2023 · f590640 · f590640
1 parent 90c6518
commit f590640
Show file tree

Hide file tree

Showing 4 changed files with 10 additions and 12 deletions.
diff --git a/.../vdk-plugins/vdk-jupyter/vdk-jupyterlab-extension/src/components/ConvertJobToNotebook.tsx b/.../vdk-plugins/vdk-jupyter/vdk-jupyterlab-extension/src/components/ConvertJobToNotebook.tsx
@@ -267,8 +267,7 @@ export const populateNotebook = async (notebookTracker: INotebookTracker) => {
         '        <li> The below cells are automatically generated corresponding to a step (.sql file or .py file with VDK run(job_input) function) in your original job.</li>\n',
         '        <li> When you see a title saying <b>"Step generated from: sample.py"</b> before some blocks of code, \n',
         '        it means that the code below that title was created from the "sample.py" file.</li>\n',
-        '        <li> Similarly, if you come across code cells that have the format <b>"job_input.execute_query(query_string)"</b> ,\n',
-        '        it means that those cells contain code generated from ".sql" files.</li>\n',
+        '        <li> Similarly, if you come across code cells that start <b>%%vdksql</b> contain code generated from ".sql" steps.</li>\n',
         '        <li> On the other hand, code cells originating from ".py" files remain unchanged.\n',
         '        However, an additional cell is included that calls the "run" function using the command <b>"run(job_input)"</b> . \n',
         '            This cell executes the "run" function from the code generated from the ".py" file.</li>\n',

diff --git a/.../vdk-plugins/vdk-jupyter/vdk-jupyterlab-extension/vdk_jupyterlab_extension/convert_job.py b/.../vdk-plugins/vdk-jupyter/vdk-jupyterlab-extension/vdk_jupyterlab_extension/convert_job.py
@@ -93,7 +93,7 @@ def _process_python_files(self, file):
     def _process_sql_files(self, file):
         with open(file) as f:
             content = f.read()
-        self._code_structure.append(f'job_input.execute_query("""{content}""")')
+        self._code_structure.append(f"%%vdksql\n{content}")
         self._removed_files.append(os.path.basename(file))
         os.remove(file)
 

diff --git a/...er/vdk-jupyterlab-extension/vdk_jupyterlab_extension/jupyter_sample_job/10_notebook.ipynb b/...er/vdk-jupyterlab-extension/vdk_jupyterlab_extension/jupyter_sample_job/10_notebook.ipynb
@@ -38,16 +38,15 @@
    },
    "outputs": [],
    "source": [
-    "\"\"\"\n",
-    "This and the subsequent cells are marked as VDK cells.\n",
+    "%%vdksql\n",
+    "/*\n",
+    "Cells with '%%vdksql' magic are VDK SQL Cell (step).\n",
     "\n",
-    "Cells can be marked or unmarked as VDK cells through the gearwheel menu\n",
-    "to the right.\n",
-    "\"\"\"\n",
+    "Note that this and the subsequent cells are marked as VDK cells.\n",
+    "Cells can be marked or unmarked as VDK cells through the gearwheel menu to the right.\n",
+    "*/\n",
     "\n",
-    "job_input.execute_query(\n",
-    "    \"CREATE TABLE IF NOT EXISTS hello_world (id NVARCHAR);\"\n",
-    ")"
+    "CREATE TABLE IF NOT EXISTS hello_world (id NVARCHAR)"
    ]
   },
   {

diff --git a/...pyterlab-extension/vdk_jupyterlab_extension/tests/test_convert_job_directory_processor.py b/...pyterlab-extension/vdk_jupyterlab_extension/tests/test_convert_job_directory_processor.py
@@ -93,7 +93,7 @@ def test_cleanup(self):
     def test_get_code_structure(self):
         self.processor.process_files()
         expected_code_structure = [
-            f'job_input.execute_query("""{self.sql_content}""")',
+            f"%%vdksql\n{self.sql_content}",
             self.py_content_run,
             self.py_content_run_multiline,
             self.py_content_run_spaces,