Translated latest questions_gen

data/questions_gen_postgres.csv

@@ -28,7 +28,7 @@ What are the publications written by authors from the 'Sociology' domain and pre
 What's the average predicted time to graduation since admission in no. of days?,SELECT avg(predicted_graduation_semester - admit_term) AS average_predicted_time_to_graduation FROM student;,advising,date_functions,
 How many students were predicted to graduate in the last 10 years?,"SELECT count(*) AS num_students_graduated FROM student WHERE predicted_graduation_semester >= DATE_TRUNC('year', CURRENT_DATE) - interval '10 year';",advising,date_functions,
 How long has it been in days since the last admitted student?,SELECT CURRENT_DATE - max(admit_term) AS duration_since_last_admitted_student FROM student;,advising,date_functions,
-Return the course id's that are offered in either semesters 1 or 2 and ends before 1pm and had an instructor on thursday,"SELECT DISTINCT co.course_id FROM public.course_offering co JOIN public.offering_instructor oi ON co.offering_id = oi.offering_id WHERE (co.semester = 1 OR co.semester = 2) AND co.end_time < '13:00:00' AND co.thursday IS NOT NULL;",advising,date_functions,
+Return the course id's that are offered in either semesters 1 or 2 and ends before 1pm and had an instructor on thursday,"SELECT DISTINCT co.course_id FROM course_offering co JOIN offering_instructor oi ON co.offering_id = oi.offering_id WHERE (co.semester = 1 OR co.semester = 2) AND co.end_time < '13:00:00' AND co.thursday IS NOT NULL;",advising,date_functions,
 What is the total number of students who found the instructor to be hilarious per course id?,"SELECT course_tags_count.course_id, SUM(course_tags_count.hilarious) AS total_hilarious FROM course_tags_count GROUP BY course_tags_count.course_id;",advising,group_by,
 What is the average clarity score for each instructor who taught a course?,"SELECT {i.name, i.instructor_id}, AVG(c.clarity_score) FROM course c JOIN course_offering co ON c.course_id = co.course_id JOIN offering_instructor oi ON co.offering_id = oi.offering_id JOIN instructor i ON oi.instructor_id = i.instructor_id GROUP BY {};",advising,group_by,
 How many course offerings have a final exam and how many do not?,"SELECT course_offering.has_final_exam, COUNT(offering_id) AS num_courses FROM course_offering GROUP BY course_offering.has_final_exam;SELECT COUNT(CASE WHEN co.has_final_exam THEN 1 END) AS num_with_final_exam, COUNT(CASE WHEN NOT co.has_final_exam THEN 1 END) AS num_without_final_exam FROM course_offering co;",advising,group_by,

translate_sql_dialect.py

@@ -45,6 +45,9 @@
 df["valid"] = ""
 df["err_msg"] = ""
 
+# fill na with empty string
+df.fillna("", inplace=True)
+
 # create db_type col where if "Snowflake" in file name, db_type = "snowflake", else db_type = "postgres"
 if "snowflake" in dataset_file:
     df["db_type"] = "snowflake"

utils/dialects.py

@@ -736,7 +736,11 @@ def create_sqlite_db(db_name, table_metadata_string_test, row_idx):
     try:
         conn = sqlite3.connect(f"{test_db_name}.db")
         cursor = conn.cursor()
-        for table in table_metadata_string_test.split("\n"):
+        for table in table_metadata_string_test.split(");"):
+            if table.strip() == "":
+                continue
+            if not table.endswith(");"):
+                table += ");"
             cursor.execute(table)
         # print(f"Tables for `{test_db_name}` created successfully")
     except Exception as err:
@@ -989,6 +993,8 @@ def test_valid_md_tsql(sql_test_list, db_name, table_metadata_string_test, row_i
         validity_added = False
         while tries < 3 and not validity_added:
             try:
+                import pyodbc
+
                 conn = pyodbc.connect(
                     f"DRIVER={creds['tsql']['driver']};SERVER={creds['tsql']['server']};DATABASE={test_db};UID={creds['tsql']['user']};PWD={creds['tsql']['password']}"
                 )