docs: add PG 42703 | PG 28P01 | MySQL 1054 (#1052)

* add PG 42703 |  PG 28P01 | MySQL 1054

* Update content/reference/mysql/error/1054-unknown-column-in-field-list.md

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Update content/reference/postgres/error/42703-undefined-column-postgres.md

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Update content/reference/postgres/error/28p01-invalid-password-postgres.md

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

* Update content/reference/mysql/error/1054-unknown-column-in-field-list.md

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>

---------

Co-authored-by: Copilot <175728472+Copilot@users.noreply.github.com>
Co-authored-by: Tianzhou <t@bytebase.com>

Author: 3 people

content/reference/mysql/error/1054-unknown-column-in-field-list.md

@@ -0,0 +1,129 @@
+---
+title: "How to Fix MySQL Error 1054: Unknown Column in 'field list'"
+---
+
+## Error Message
+
+```sql
+ERROR 1054 (42S22): Unknown column 'emial' in 'field list'
+```
+
+The error can also appear with different contexts:
+
+```sql
+ERROR 1054 (42S22): Unknown column 'status' in 'where clause'
+ERROR 1054 (42S22): Unknown column 'u.name' in 'on clause'
+ERROR 1054 (42S22): Unknown column 'total' in 'order clause'
+```
+
+## Description
+
+MySQL raises error 1054 when a query references a column that does not exist in the specified table or expression. The SQLSTATE code is 42S22. This is the MySQL counterpart to [PostgreSQL error 42703](/reference/postgres/error/42703-undefined-column-postgres). The error message helpfully tells you which clause contains the bad reference — `field list` (SELECT), `where clause`, `on clause`, or `order clause`.
+
+## Common Causes
+
+1. **Typo in the column name**: `SELECT emial FROM users` when the column is `email`
+2. **Missing or wrong table alias**: Referencing a column without qualifying which table it belongs to in a JOIN
+3. **Column alias used in WHERE or HAVING**: MySQL doesn't allow referencing a SELECT alias in the WHERE clause; aliases are only usable in ORDER BY (and sometimes HAVING)
+4. **Column dropped or renamed**: A migration changed the column but application queries still use the old name
+5. **Wrong table referenced**: The column exists in a different table than expected
+6. **Backtick quoting issues**: Reserved words used as column names without backticks
+7. **Subquery column not visible**: Referencing a column from an inner query that isn't exposed to the outer query
+
+## How to Fix
+
+### Solution 1: Check the Actual Column Names
+
+```sql
+-- List all columns in a table
+DESCRIBE users;
+
+-- Or use INFORMATION_SCHEMA
+SELECT COLUMN_NAME, DATA_TYPE
+FROM INFORMATION_SCHEMA.COLUMNS
+WHERE TABLE_SCHEMA = DATABASE() AND TABLE_NAME = 'users'
+ORDER BY ORDINAL_POSITION;
+```
+
+### Solution 2: Qualify Columns with Table Aliases in JOINs
+
+```sql
+-- Bad: ambiguous column reference
+SELECT name FROM orders JOIN customers ON orders.customer_id = customers.id;
+
+-- Good: use table alias
+SELECT c.name FROM orders o JOIN customers c ON o.customer_id = c.id;
+```
+
+### Solution 3: Don't Use Aliases in WHERE — Repeat the Expression
+
+```sql
+-- Bad: alias not visible in WHERE
+SELECT CONCAT(first_name, ' ', last_name) AS full_name
+FROM users WHERE full_name LIKE '%Smith%';
+
+-- Good: repeat the expression
+SELECT CONCAT(first_name, ' ', last_name) AS full_name
+FROM users WHERE CONCAT(first_name, ' ', last_name) LIKE '%Smith%';
+
+-- Alternative: use a subquery
+SELECT * FROM (
+  SELECT *, CONCAT(first_name, ' ', last_name) AS full_name FROM users
+) t WHERE full_name LIKE '%Smith%';
+```
+
+### Solution 4: Quote Reserved Words with Backticks
+
+```sql
+-- Bad: 'order' is a reserved word
+SELECT order FROM purchases;
+
+-- Good: backtick the reserved word
+SELECT `order` FROM purchases;
+```
+
+### Solution 5: Fix ON Clause References
+
+When the error points to `'on clause'`, the column reference in a JOIN condition is wrong:
+
+```sql
+-- Bad: wrong column name in ON clause
+SELECT * FROM orders o JOIN customers c ON o.cust_id = c.id;
+
+-- Good: use the correct column name
+SELECT * FROM orders o JOIN customers c ON o.customer_id = c.id;
+```
+
+### Solution 6: Verify After Migrations
+
+If a column was recently renamed or dropped:
+
+```sql
+-- Check if the old column still exists
+SHOW COLUMNS FROM users LIKE 'user_name';
+
+-- Check what columns currently exist
+SHOW COLUMNS FROM users;
+```
+
+## Common scenarios
+
+**With `SELECT *` in views:** If a view was created with `SELECT *` and the underlying table later dropped a column, queries against the view will fail with 1054. Recreate the view after schema changes.
+
+**In INSERT statements:** `INSERT INTO users (emial) VALUES ('a@b.com')` fails if the column is `email`. The error message will say `Unknown column 'emial' in 'field list'`.
+
+**In ORDER BY with UNION:** When using UNION, ORDER BY can only reference columns from the first SELECT or use positional numbers:
+
+```sql
+-- Bad
+SELECT id, email FROM users UNION SELECT id, address FROM contacts ORDER BY email;
+
+-- Good: use column position
+SELECT id, email FROM users UNION SELECT id, address FROM contacts ORDER BY 2;
+```
+
+<HintBlock type="info">
+
+Bytebase's [SQL Review](https://www.bytebase.com/docs/sql-review/review-rules/) can catch unknown column references during change review before they reach production. See also [ERROR 1146: Table Doesn't Exist](/reference/mysql/error/1146-table-doesnt-exist) for the related table-not-found error.
+
+</HintBlock>

content/reference/mysql/error/_layout.md

@@ -31,6 +31,8 @@
 
 ## Query Syntax & Logic
 
+### [ERROR 1054 (42S22): Unknown column in 'field list'](/reference/mysql/error/1054-unknown-column-in-field-list)
+
 ### [ERROR 1064 (42000): You have an error in your SQL syntax](/reference/mysql/error/1064-syntax-error)
 
 ### [ERROR 1055 (42000): SELECT list is not in GROUP BY clause and contains nonaggregated column](/reference/mysql/error/1055-select-list-not-in-group-by)

content/reference/postgres/error/28p01-invalid-password-postgres.md

@@ -0,0 +1,143 @@
+---
+title: 'ERROR 28P01: Password Authentication Failed in Postgres'
+---
+
+## Error Message
+
+```sql
+FATAL: password authentication failed for user "app_user"
+SQLSTATE: 28P01
+```
+
+You may also see a related log entry on the server:
+
+```
+DETAIL: User "app_user" has no password assigned.
+```
+
+Or when connecting via `psql` or a driver:
+
+```
+psql: error: connection to server at "localhost" (127.0.0.1), port 5432 failed:
+FATAL: password authentication failed for user "app_user"
+```
+
+## Description
+
+PostgreSQL raises error 28P01 when a client provides a password that does not match the stored credential for the specified user, or when password authentication is required but no valid password is set. The full SQLSTATE code is 28P01 (`invalid_password`). This is the most common connection error for PostgreSQL and is almost always a configuration issue — not a bug in your application.
+
+## Causes
+
+- **Wrong password.** The most obvious cause — the password supplied doesn't match what's stored for that role.
+- **Role has no password set.** The role was created with `CREATE ROLE app_user LOGIN` but without a `PASSWORD` clause, so any password attempt fails.
+- **Wrong user.** Connecting as `postgres` when the application should connect as `app_user`, or vice versa.
+- **pg_hba.conf requires password auth but the role uses a different method.** The `pg_hba.conf` entry says `md5` or `scram-sha-256` but the role's password was set with a different encryption method.
+- **Password contains special characters.** Characters like `@`, `#`, `%` in the password may be misinterpreted by connection strings if not properly escaped or quoted.
+- **Environment variable or config file has stale credentials.** The password was rotated but `.pgpass`, `PGPASSWORD`, or the application config still has the old one.
+- **Connecting to the wrong server.** The credentials are valid on staging but you're connecting to production (or vice versa).
+- **SCRAM vs MD5 mismatch.** PostgreSQL 14+ defaults to `scram-sha-256`. If the password was stored as MD5 and `pg_hba.conf` requires SCRAM, authentication fails.
+
+## Solutions
+
+1. **Reset the password:**
+
+   ```sql
+   -- Connect as a superuser
+   ALTER ROLE app_user WITH PASSWORD 'new_secure_password';
+   ```
+
+2. **Verify the role exists and has LOGIN privilege:**
+
+   ```sql
+   SELECT rolname, rolcanlogin
+   FROM pg_roles
+   WHERE rolname = 'app_user';
+   ```
+
+   If `rolcanlogin` is `f`, grant login:
+
+   ```sql
+   ALTER ROLE app_user WITH LOGIN;
+   ```
+
+3. **Check pg_hba.conf authentication method:**
+
+   ```bash
+   # Find the active pg_hba.conf
+   psql -U postgres -c "SHOW hba_file;"
+   ```
+
+   Look for the line matching your connection. Common entries:
+
+   ```
+   # TYPE  DATABASE  USER       ADDRESS        METHOD
+   host    all       all        127.0.0.1/32   scram-sha-256
+   host    all       all        0.0.0.0/0      md5
+   ```
+
+   After editing, reload:
+
+   ```sql
+   SELECT pg_reload_conf();
+   ```
+
+4. **Fix SCRAM vs MD5 mismatch:**
+
+   ```sql
+   -- Check current password encryption setting
+   SHOW password_encryption;
+
+   -- If it shows 'scram-sha-256' but pg_hba.conf says 'md5', either:
+   -- Option A: Change pg_hba.conf to scram-sha-256 (recommended)
+   -- Option B: Set encryption to md5 and reset the password
+   SET password_encryption = 'md5';
+   ALTER ROLE app_user WITH PASSWORD 'the_password';
+   ```
+
+5. **Check .pgpass or connection string:**
+
+   ```bash
+   # .pgpass format: hostname:port:database:username:password
+   cat ~/.pgpass
+
+   # Ensure correct permissions
+   chmod 600 ~/.pgpass
+   ```
+
+   For connection strings, escape special characters:
+
+   ```
+   postgresql://app_user:p%40ssw0rd@localhost:5432/mydb
+   ```
+
+   `@` → `%40`, `#` → `%23`, `%` → `%25`
+
+6. **Verify you're connecting to the right server:**
+
+   ```sql
+   -- After connecting, check where you are
+   SELECT inet_server_addr(), inet_server_port(), current_user, current_database();
+   ```
+
+7. **Check server logs for more detail:**
+
+   ```bash
+   # The server log often has a DETAIL line explaining what went wrong
+   tail -20 /var/log/postgresql/postgresql-16-main.log
+   ```
+
+## Common scenarios
+
+**In Docker and containers:** The `POSTGRES_PASSWORD` environment variable only sets the superuser password on first initialization. If you change it in `docker-compose.yml` and restart without deleting the volume, the old password persists. Either drop the volume (`docker volume rm ...`) or connect and run `ALTER ROLE`.
+
+**After PostgreSQL major version upgrade:** PostgreSQL 14 changed the default `password_encryption` from `md5` to `scram-sha-256`. After upgrading, existing MD5 passwords still work if `pg_hba.conf` uses `md5`, but if you change `pg_hba.conf` to `scram-sha-256`, all users must reset their passwords.
+
+**In connection pools (PgBouncer, Pgpool-II):** The pooler authenticates users independently. If the database password is changed but the pooler's `userlist.txt` or auth config isn't updated, connections through the pool fail with 28P01 even though direct connections work.
+
+**With cloud-managed databases (RDS, Cloud SQL, Azure):** Cloud providers manage `pg_hba.conf` internally. If you get 28P01 on RDS, verify the password via the AWS console or `aws rds modify-db-instance`. For Cloud SQL, use `gcloud sql users set-password`.
+
+<HintBlock type="info">
+
+Bytebase provides a [secure credential management workflow](https://www.bytebase.com/docs/security/secret/) that avoids sharing database passwords directly. For more on PostgreSQL access control, see [Database Access Control Best Practices](/blog/database-access-control-best-practices/).
+
+</HintBlock>

content/reference/postgres/error/42703-undefined-column-postgres.md

@@ -0,0 +1,108 @@
+---
+title: 'ERROR 42703: Column Does Not Exist in Postgres'
+---
+
+## Error Message
+
+```sql
+ERROR: column "emial" does not exist
+LINE 1: SELECT emial FROM users;
+               ^
+SQLSTATE: 42703
+```
+
+## Description
+
+PostgreSQL raises error 42703 when a query references a column that the database cannot find in the target table or expression. The full SQLSTATE code is 42703 (`undefined_column`). This is one of the most common PostgreSQL errors — closely related to [42P01 (undefined table)](/reference/postgres/error/42p01-undefined-table-postgres) — and usually traces back to a spelling mistake, a missing alias, or a schema mismatch.
+
+## Causes
+
+- **Typo in the column name.** `SELECT emial FROM users` fails because the column is actually `email`.
+- **Missing table alias or wrong alias.** In a multi-table query, referencing `SELECT name FROM orders JOIN customers c ON ...` fails if `name` exists in `customers` but not `orders` and no alias is specified.
+- **Case sensitivity.** If a column was created with double quotes (`"Status"`), you must always reference it as `"Status"`. Without quotes, PostgreSQL folds identifiers to lowercase, so `SELECT Status` becomes `SELECT status` which won't match `"Status"`.
+- **Column dropped or renamed.** A migration renamed `user_name` to `username` but application code still references the old name.
+- **Wrong table in a JOIN.** The column exists in a different table than the one being referenced.
+- **Subquery or CTE column not exposed.** A subquery selects `id` but the outer query references `user_id` from that subquery.
+- **Schema mismatch after migration.** The application expects a column that a migration hasn't created yet, or that was rolled back.
+
+## Solutions
+
+1. **Check the actual column names in the table:**
+
+   ```sql
+   SELECT column_name, data_type
+   FROM information_schema.columns
+   WHERE table_name = 'users'
+   ORDER BY ordinal_position;
+   ```
+
+2. **Use a table alias to disambiguate columns in JOINs:**
+
+   ```sql
+   -- Bad: ambiguous or wrong table
+   SELECT name FROM orders JOIN customers ON orders.customer_id = customers.id;
+
+   -- Good: qualify with alias
+   SELECT c.name FROM orders o JOIN customers c ON o.customer_id = c.id;
+   ```
+
+3. **Handle case-sensitive column names:**
+
+   ```sql
+   -- If created with double quotes
+   SELECT "Status" FROM orders;
+
+   -- Check actual casing
+   SELECT column_name FROM information_schema.columns
+   WHERE table_name = 'orders' AND column_name ILIKE '%status%';
+   ```
+
+4. **Verify the column in psql:**
+
+   ```sql
+   \d users
+   ```
+
+5. **Check subquery or CTE column names:**
+
+   ```sql
+   -- Bad: outer query references a column the CTE doesn't expose
+   WITH active AS (SELECT id FROM users WHERE active = true)
+   SELECT user_id FROM active;
+
+   -- Good: match the column name
+   WITH active AS (SELECT id FROM users WHERE active = true)
+   SELECT id FROM active;
+   ```
+
+6. **Run pending migrations:**
+
+   If the column should have been added by a migration, verify it was applied:
+
+   ```bash
+   # Check migration status
+   flyway info
+   # Or check the change history in Bytebase
+   ```
+
+## Common scenarios
+
+**In ORMs and application code:** ORMs map model fields to column names. If you rename a field in your model but forget to generate a migration, or if the migration hasn't been applied, PostgreSQL will report 42703 at runtime. Check `\d tablename` against your model definition.
+
+**In GROUP BY and ORDER BY:** PostgreSQL doesn't allow referencing column aliases in `WHERE` or `HAVING` (though `ORDER BY` is fine). `SELECT email AS e FROM users WHERE e LIKE '%@%'` fails with 42703 because aliases aren't visible in `WHERE`.
+
+```sql
+-- Bad
+SELECT email AS e FROM users WHERE e LIKE '%@%';
+
+-- Good
+SELECT email AS e FROM users WHERE email LIKE '%@%';
+```
+
+**After ALTER TABLE:** If a column was renamed with `ALTER TABLE users RENAME COLUMN user_name TO username`, any view, function, or application query still using the old name will fail with 42703. Search your codebase for the old column name.
+
+<HintBlock type="info">
+
+Bytebase's [SQL Review](https://www.bytebase.com/docs/sql-review/review-rules/) can catch references to non-existent columns during change review, before they reach production. See also [ERROR 42P01: Relation Does Not Exist](/reference/postgres/error/42p01-undefined-table-postgres) for the related table-not-found error.
+
+</HintBlock>