-- Simple Fix for Malformed Time Entries
-- Fixes times like "34 AM" (missing hour) by adding a default hour
-- Run this in your PostgreSQL database

-- Step 1: Check for malformed entries
SELECT 
  "CreatedDate",
  "CreatedTime",
  COUNT(*) as count
FROM "OptionsFlow_monthly"
WHERE "CreatedTime" IS NOT NULL
  AND "CreatedTime"::text ~ '^\d{2}\s*(AM|PM)$'  -- Two digits + AM/PM (likely minute only)
  AND "CreatedTime"::text !~ '^\d{1,2}:\d{2}'     -- No colon (missing hour)
GROUP BY "CreatedDate", "CreatedTime"
ORDER BY count DESC
LIMIT 50;

-- Step 2: Fix entries where time is just "XX AM" or "XX PM" (missing hour)
-- Strategy: If the number is > 12, it's definitely a minute, so add hour 9 (AM) or 1 (PM)
-- If number <= 12, check context or default to 9 AM / 1 PM

UPDATE "OptionsFlow_monthly"
SET "CreatedTime" = 
  CASE 
    -- Pattern: "34 AM" -> "9:34 AM" (number > 12, must be minutes)
    WHEN "CreatedTime"::text ~ '^(\d{2})\s*AM$' 
      AND (REGEXP_REPLACE("CreatedTime"::text, '^(\d{2})\s*AM$', '\1'))::int > 12
    THEN 
      '9:' || REGEXP_REPLACE("CreatedTime"::text, '^(\d{2})\s*AM$', '\1') || ':00 AM'
    
    -- Pattern: "34 PM" -> "1:34 PM" (number > 12, must be minutes)
    WHEN "CreatedTime"::text ~ '^(\d{2})\s*PM$' 
      AND (REGEXP_REPLACE("CreatedTime"::text, '^(\d{2})\s*PM$', '\1'))::int > 12
    THEN 
      '1:' || REGEXP_REPLACE("CreatedTime"::text, '^(\d{2})\s*PM$', '\1') || ':00 PM'
    
    -- Pattern: "12 AM" or less -> "9:XX AM" (could be hour or minute, default to 9 AM)
    WHEN "CreatedTime"::text ~ '^(\d{1,2})\s*AM$' 
      AND (REGEXP_REPLACE("CreatedTime"::text, '^(\d{1,2})\s*AM$', '\1'))::int <= 12
    THEN 
      '9:' || LPAD(REGEXP_REPLACE("CreatedTime"::text, '^(\d{1,2})\s*AM$', '\1'), 2, '0') || ':00 AM'
    
    -- Pattern: "12 PM" or less -> "1:XX PM" (could be hour or minute, default to 1 PM)
    WHEN "CreatedTime"::text ~ '^(\d{1,2})\s*PM$' 
      AND (REGEXP_REPLACE("CreatedTime"::text, '^(\d{1,2})\s*PM$', '\1'))::int <= 12
    THEN 
      '1:' || LPAD(REGEXP_REPLACE("CreatedTime"::text, '^(\d{1,2})\s*PM$', '\1'), 2, '0') || ':00 PM'
    
    -- Pattern: ":34 AM" (missing hour, has colon) -> "9:34 AM"
    WHEN "CreatedTime"::text ~ '^:(\d{1,2})\s*AM$'
    THEN 
      '9:' || LPAD(REGEXP_REPLACE("CreatedTime"::text, '^:(\d{1,2})\s*AM$', '\1'), 2, '0') || ':00 AM'
    
    -- Pattern: ":34 PM" (missing hour, has colon) -> "1:34 PM"
    WHEN "CreatedTime"::text ~ '^:(\d{1,2})\s*PM$'
    THEN 
      '1:' || LPAD(REGEXP_REPLACE("CreatedTime"::text, '^:(\d{1,2})\s*PM$', '\1'), 2, '0') || ':00 PM'
    
    ELSE "CreatedTime"  -- Keep as-is if doesn't match patterns
  END
WHERE "CreatedTime" IS NOT NULL
  AND (
    -- Match patterns: "XX AM/PM" without colon, or ":XX AM/PM"
    ("CreatedTime"::text ~ '^\d{1,2}\s*(AM|PM)$' AND "CreatedTime"::text !~ '^\d{1,2}:\d{2}')
    OR "CreatedTime"::text ~ '^:\d{1,2}\s*(AM|PM)$'
  );

-- Step 3: Verify the fix
SELECT 
  COUNT(*) as remaining_malformed,
  'Remaining entries with malformed time format' as description
FROM "OptionsFlow_monthly"
WHERE "CreatedTime" IS NOT NULL
  AND (
    ("CreatedTime"::text ~ '^\d{1,2}\s*(AM|PM)$' AND "CreatedTime"::text !~ '^\d{1,2}:\d{2}')
    OR "CreatedTime"::text ~ '^:\d{1,2}\s*(AM|PM)$'
  );

-- Step 4: Show sample of fixed entries
SELECT 
  "CreatedDate",
  "CreatedTime",
  "Symbol",
  COUNT(*) as count
FROM "OptionsFlow_monthly"
WHERE "CreatedTime" IS NOT NULL
  AND "CreatedTime"::text ~ '^\d{1,2}:\d{2}:\d{2}\s*(AM|PM)$'  -- Properly formatted now
GROUP BY "CreatedDate", "CreatedTime", "Symbol"
ORDER BY "CreatedDate" DESC
LIMIT 20;