[Bug] Training Instance 3153, gold query refers to a table that doesn't exist in the db.

[hXtreme]

Training Instance 3153:

{
  "db_id": "assets_maintenance",
  "query": "SELECT T1.company_name FROM Third_Party_Companies AS T1 JOIN Maintenance_Contracts AS T2 ON T1.company_id  =  T2.maintenance_contract_company_id JOIN Ref_Company_Types AS T3 ON T1.company_type_code  =  T3.company_type_code ORDER BY T2.contract_end_date DESC LIMIT 1",
  "query_toks": [...],
  "query_toks_no_value": [...],
  "question": "What is the description of the type of the company who concluded its contracts most recently?",
  "question_toks": [...],
  "sql": {...}
}

The table Ref_Company_Types doesn't exist in assets_maintenance db.

assets_maintenance schema:

{
  "db_id": "assets_maintenance",
  "table_names": [
    "third party companies",
    "maintenance contracts",
    "parts",
    "skills",
    "staff",
    "assets",
    "asset parts",
    "maintenance engineers",
    "engineer skills",
    "fault log",
    "engineer visits",
    "part faults",
    "fault log parts",
    "skills required to fix"
  ],
  "table_names_original": [
    "Third_Party_Companies",
    "Maintenance_Contracts",
    "Parts",
    "Skills",
    "Staff",
    "Assets",
    "Asset_Parts",
    "Maintenance_Engineers",
    "Engineer_Skills",
    "Fault_Log",
    "Engineer_Visits",
    "Part_Faults",
    "Fault_Log_Parts",
    "Skills_Required_To_Fix"
  ],
  "column_names": [...],
  "column_names_original": [...],
  "column_types": [...],
  "primary_keys": [...],
  "foreign_keys": [...]
}

[hXtreme]

Related to #51 #52

[ericzakariasson]

Hey @hXtreme, did you ever get this to work? We're running into the same issue

[hXtreme]

@ericzakariasson I'd recommend just dropping that one example.

[ericzakariasson]

@hXtreme Thanks, that's what we ended up doing! All in all we skipped 3 questions/queries due them not working for different reasons

[hXtreme]

@ericzakariasson might be nice to share some information about the ones you had to drop for others.

tbh @taoyds should be taking charge in fixing issues with the dataset they've published.