386 lines
12 KiB
Markdown
386 lines
12 KiB
Markdown
|
```sql
|
||
|
|
|
||
|
|
SELECT
|
||
|
|
column_name,
|
||
|
|
comments
|
||
|
|
FROM
|
||
|
|
all_col_comments
|
||
|
|
WHERE
|
||
|
|
owner = 'ZC'
|
||
|
|
AND table_name = 'BASE_USER';
|
||
|
|
```
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
To close all sessions associated with the user `'zc'` in an Oracle database, you need to identify those sessions and then terminate them. This process requires administrative privileges. Below are the steps to accomplish this:
|
||
|
|
|
||
|
|
1. **Connect to the Database as an Administrator**
|
||
|
|
|
||
|
|
Open SQL*Plus or your preferred SQL client and connect as a user with administrative privileges (such as `SYSDBA`):
|
||
|
|
|
||
|
|
```sql
|
||
|
|
sqlplus / as sysdba
|
||
|
|
```
|
||
|
|
|
||
|
|
2. **Identify the Sessions Belonging to User 'zc'**
|
||
|
|
|
||
|
|
Execute the following query to retrieve the `SID` and `SERIAL#` of all sessions for user `'zc'`:
|
||
|
|
|
||
|
|
```sql
|
||
|
|
SELECT SID, SERIAL# FROM V$SESSION WHERE USERNAME = 'ZC';
|
||
|
|
```
|
||
|
|
|
||
|
|
*Note*: Oracle usernames are usually stored in uppercase unless created with double quotes.
|
||
|
|
|
||
|
|
3. **Terminate the Sessions**
|
||
|
|
|
||
|
|
You can terminate each session individually using the `ALTER SYSTEM KILL SESSION` command:
|
||
|
|
|
||
|
|
```sql
|
||
|
|
ALTER SYSTEM KILL SESSION 'SID,SERIAL#';
|
||
|
|
```
|
||
|
|
|
||
|
|
Replace `SID` and `SERIAL#` with the values obtained from the previous query.
|
||
|
|
|
||
|
|
**Example**:
|
||
|
|
|
||
|
|
If the query returns `SID` = 123 and `SERIAL#` = 4567, execute:
|
||
|
|
|
||
|
|
```sql
|
||
|
|
ALTER SYSTEM KILL SESSION '123,4567';
|
||
|
|
```
|
||
|
|
|
||
|
|
4. **Automate the Termination of All Sessions**
|
||
|
|
|
||
|
|
To kill all sessions for user `'zc'` in one step, you can use a PL/SQL block:
|
||
|
|
|
||
|
|
```sql
|
||
|
|
BEGIN
|
||
|
|
FOR rec IN (SELECT SID, SERIAL# FROM V$SESSION WHERE USERNAME = 'ZC') LOOP
|
||
|
|
EXECUTE IMMEDIATE 'ALTER SYSTEM KILL SESSION ''' || rec.SID || ',' || rec.SERIAL# || ''' IMMEDIATE';
|
||
|
|
END LOOP;
|
||
|
|
END;
|
||
|
|
/
|
||
|
|
```
|
||
|
|
|
||
|
|
This script loops through each session and issues the kill command.
|
||
|
|
|
||
|
|
5. **Verify the Sessions Are Closed**
|
||
|
|
|
||
|
|
Run the initial query again to ensure all sessions have been terminated:
|
||
|
|
|
||
|
|
```sql
|
||
|
|
SELECT SID, SERIAL# FROM V$SESSION WHERE USERNAME = 'ZC';
|
||
|
|
```
|
||
|
|
|
||
|
|
If no rows are returned, all sessions have been successfully closed.
|
||
|
|
|
||
|
|
**Important Considerations**:
|
||
|
|
|
||
|
|
- **Transactions**: Killing a session that is in the middle of a transaction can cause that transaction to be rolled back. Ensure that terminating these sessions won't negatively impact your database integrity.
|
||
|
|
- **Privileges**: You must have the necessary privileges (`ALTER SYSTEM` privilege) to kill sessions.
|
||
|
|
- **Immediate Termination**: Adding the `IMMEDIATE` clause forces the session to terminate without waiting for ongoing transactions to complete.
|
||
|
|
- **Oracle Versions**: In Oracle 12c and newer, you can use `ALTER SYSTEM DISCONNECT SESSION` for immediate termination:
|
||
|
|
|
||
|
|
```sql
|
||
|
|
ALTER SYSTEM DISCONNECT SESSION 'SID,SERIAL#' IMMEDIATE;
|
||
|
|
```
|
||
|
|
|
||
|
|
By following these steps, you can close all sessions associated with the user `'zc'` in your Oracle database.
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
|
||
|
|
When the source and target databases have different character sets (NLS settings), exporting and importing data in Oracle requires careful handling to ensure data integrity. Incorrect handling can lead to character corruption, loss of data, or import errors. Below, I'll guide you through the steps to perform `exp` (export) and `imp` (import) when the source and target `NLS_LANG` settings are not the same.
|
||
|
|
|
||
|
|
---
|
||
|
|
|
||
|
|
### **Understanding NLS_LANG and Character Sets**
|
||
|
|
|
||
|
|
- **`NLS_LANG` Environment Variable**: Determines the language, territory, and character set settings for Oracle client applications.
|
||
|
|
- **Character Set Conversion**: Oracle performs character set conversion between the client (as specified by `NLS_LANG`) and the database character set.
|
||
|
|
|
||
|
|
**Key Principle**: To prevent unwanted character set conversion and potential data corruption, set `NLS_LANG` to match the **database character set** when using `exp` and `imp` utilities.
|
||
|
|
|
||
|
|
---
|
||
|
|
|
||
|
|
### **Steps for Exporting and Importing Between Different NLS_LANG Settings**
|
||
|
|
|
||
|
|
#### **1. Determine the Character Sets**
|
||
|
|
|
||
|
|
**Source Database**:
|
||
|
|
|
||
|
|
```sql
|
||
|
|
SELECT value FROM nls_database_parameters WHERE parameter = 'NLS_CHARACTERSET';
|
||
|
|
```
|
||
|
|
|
||
|
|
**Target Database**:
|
||
|
|
|
||
|
|
```sql
|
||
|
|
SELECT value FROM nls_database_parameters WHERE parameter = 'NLS_CHARACTERSET';
|
||
|
|
```
|
||
|
|
|
||
|
|
#### **2. Set NLS_LANG for Export**
|
||
|
|
|
||
|
|
- **Objective**: Prevent character set conversion during export.
|
||
|
|
- **Action**: Set `NLS_LANG` to match the **source database character set**.
|
||
|
|
|
||
|
|
**Example**:
|
||
|
|
|
||
|
|
If the source database character set is `WE8MSWIN1252`:
|
||
|
|
|
||
|
|
- **Windows**:
|
||
|
|
|
||
|
|
```cmd
|
||
|
|
SET NLS_LANG=AMERICAN_AMERICA.WE8MSWIN1252
|
||
|
|
```
|
||
|
|
|
||
|
|
- **Unix/Linux**:
|
||
|
|
|
||
|
|
```bash
|
||
|
|
export NLS_LANG=AMERICAN_AMERICA.WE8MSWIN1252
|
||
|
|
```
|
||
|
|
|
||
|
|
#### **3. Perform the Export (`exp`)**
|
||
|
|
|
||
|
|
Run the `exp` command after setting `NLS_LANG`:
|
||
|
|
|
||
|
|
```cmd
|
||
|
|
exp userid=username/password@source_db file=export.dmp log=export.log [other options]
|
||
|
|
```
|
||
|
|
|
||
|
|
- **No Character Set Conversion**: Since `NLS_LANG` matches the database character set, Oracle doesn't perform character set conversion during export.
|
||
|
|
|
||
|
|
#### **4. Transfer the Dump File**
|
||
|
|
|
||
|
|
Copy `export.dmp` to the target machine if necessary.
|
||
|
|
|
||
|
|
#### **5. Set NLS_LANG for Import**
|
||
|
|
|
||
|
|
- **Objective**: Ensure correct character set conversion during import.
|
||
|
|
- **Action**: Set `NLS_LANG` to match the **target database character set**.
|
||
|
|
|
||
|
|
**Example**:
|
||
|
|
|
||
|
|
If the target database character set is `AL32UTF8`:
|
||
|
|
|
||
|
|
- **Windows**:
|
||
|
|
|
||
|
|
```cmd
|
||
|
|
SET NLS_LANG=AMERICAN_AMERICA.AL32UTF8
|
||
|
|
```
|
||
|
|
|
||
|
|
- **Unix/Linux**:
|
||
|
|
|
||
|
|
```bash
|
||
|
|
export NLS_LANG=AMERICAN_AMERICA.AL32UTF8
|
||
|
|
```
|
||
|
|
|
||
|
|
#### **6. Perform the Import (`imp`)**
|
||
|
|
|
||
|
|
Run the `imp` command:
|
||
|
|
|
||
|
|
```cmd
|
||
|
|
imp userid=username/password@target_db file=export.dmp log=import.log [other options]
|
||
|
|
```
|
||
|
|
|
||
|
|
- **Character Set Conversion**: Oracle converts character data from the dump file's character set (source database character set) to the target database character set.
|
||
|
|
|
||
|
|
#### **7. Verify the Imported Data**
|
||
|
|
|
||
|
|
After import, check the data integrity, especially for special characters or multilingual data.
|
||
|
|
|
||
|
|
---
|
||
|
|
|
||
|
|
### **Detailed Explanation**
|
||
|
|
|
||
|
|
#### **Why Set NLS_LANG to the Database Character Set?**
|
||
|
|
|
||
|
|
- **During Export**:
|
||
|
|
- Setting `NLS_LANG` to match the source database character set ensures that no character set conversion occurs between the database and the export utility.
|
||
|
|
- Data is exported exactly as it is stored in the database.
|
||
|
|
|
||
|
|
- **During Import**:
|
||
|
|
- Setting `NLS_LANG` to match the target database character set allows Oracle to correctly interpret the character data from the dump file and convert it to the target database's character set if necessary.
|
||
|
|
- This prevents character corruption due to incorrect character set conversion.
|
||
|
|
|
||
|
|
#### **Understanding Character Set Conversion**
|
||
|
|
|
||
|
|
- **Export Process**:
|
||
|
|
- **Database → Client (`exp` utility)**: If `NLS_LANG` matches the database character set, no conversion occurs.
|
||
|
|
|
||
|
|
- **Import Process**:
|
||
|
|
- **Dump File → Client (`imp` utility)**: The `imp` utility reads data in the dump file's character set.
|
||
|
|
- **Client (`imp` utility) → Target Database**: Oracle converts character data from the client character set (specified by `NLS_LANG`) to the target database character set.
|
||
|
|
|
||
|
|
---
|
||
|
|
|
||
|
|
### **Example Scenario**
|
||
|
|
|
||
|
|
**Source Database**:
|
||
|
|
|
||
|
|
- **Character Set**: `WE8ISO8859P1`
|
||
|
|
- **NLS_LANG**: `AMERICAN_AMERICA.WE8ISO8859P1`
|
||
|
|
|
||
|
|
**Target Database**:
|
||
|
|
|
||
|
|
- **Character Set**: `AL32UTF8`
|
||
|
|
- **NLS_LANG**: `AMERICAN_AMERICA.AL32UTF8`
|
||
|
|
|
||
|
|
#### **Export Steps**
|
||
|
|
|
||
|
|
1. **Set NLS_LANG**:
|
||
|
|
|
||
|
|
```cmd
|
||
|
|
SET NLS_LANG=AMERICAN_AMERICA.WE8ISO8859P1
|
||
|
|
```
|
||
|
|
|
||
|
|
2. **Run Export**:
|
||
|
|
|
||
|
|
```cmd
|
||
|
|
exp userid=source_user/password@source_db file=export.dmp log=export.log
|
||
|
|
```
|
||
|
|
|
||
|
|
#### **Import Steps**
|
||
|
|
|
||
|
|
1. **Set NLS_LANG**:
|
||
|
|
|
||
|
|
```cmd
|
||
|
|
SET NLS_LANG=AMERICAN_AMERICA.AL32UTF8
|
||
|
|
```
|
||
|
|
|
||
|
|
2. **Run Import**:
|
||
|
|
|
||
|
|
```cmd
|
||
|
|
imp userid=target_user/password@target_db file=export.dmp log=import.log
|
||
|
|
```
|
||
|
|
|
||
|
|
---
|
||
|
|
|
||
|
|
### **Additional Considerations**
|
||
|
|
|
||
|
|
#### **Character Set Compatibility**
|
||
|
|
|
||
|
|
- Ensure the target character set can represent all characters from the source character set.
|
||
|
|
- **Superset Character Sets**: If the target character set is a superset of the source, conversion should be lossless (e.g., importing from `WE8ISO8859P1` to `AL32UTF8`).
|
||
|
|
|
||
|
|
#### **Testing**
|
||
|
|
|
||
|
|
- **Test Import**: Before full-scale import, test with a small dataset to verify character data integrity.
|
||
|
|
|
||
|
|
#### **Data Pump Utilities**
|
||
|
|
|
||
|
|
- **`expdp` and `impdp`**: For Oracle 10g and later, consider using Data Pump utilities which handle character set conversions more efficiently.
|
||
|
|
- **Note**: Data Pump utilities run on the server side and generally require less attention to `NLS_LANG`, but it's still good practice to set it appropriately if your parameters include non-ASCII characters.
|
||
|
|
|
||
|
|
#### **Locale-Specific Data**
|
||
|
|
|
||
|
|
- **Date Formats, Numeric Formats**: Other NLS parameters (like `NLS_DATE_FORMAT`, `NLS_NUMERIC_CHARACTERS`) may need to be set if your data depends on locale-specific formats.
|
||
|
|
|
||
|
|
---
|
||
|
|
|
||
|
|
### **Common Mistakes to Avoid**
|
||
|
|
|
||
|
|
#### **Setting NLS_LANG to the Wrong Character Set**
|
||
|
|
|
||
|
|
- **Do Not**: Set `NLS_LANG` to the operating system character set unless it matches the database character set.
|
||
|
|
- **Result**: Misalignment can cause Oracle to perform unwanted character set conversions, leading to data corruption.
|
||
|
|
|
||
|
|
#### **Ignoring NCHAR and NVARCHAR Data**
|
||
|
|
|
||
|
|
- **Note**: `NLS_NCHAR_CHARACTERSET` may also need to be considered if your databases use NCHAR, NVARCHAR2, or NCLOB data types.
|
||
|
|
|
||
|
|
---
|
||
|
|
|
||
|
|
### **Frequently Asked Questions**
|
||
|
|
|
||
|
|
**Q1: Can I set `NLS_LANG` to the same value during both export and import?**
|
||
|
|
|
||
|
|
- **A**: While possible, it's not recommended unless the source and target databases have the same character set. Setting `NLS_LANG` to match the respective database character sets at each stage ensures proper handling.
|
||
|
|
|
||
|
|
**Q2: What if I can't match `NLS_LANG` to the database character set?**
|
||
|
|
|
||
|
|
- **A**: If you cannot set `NLS_LANG` to match the database character set, you risk character data corruption. It's crucial to find out the correct character set and set `NLS_LANG` accordingly.
|
||
|
|
|
||
|
|
**Q3: How do I handle data that includes special or non-English characters?**
|
||
|
|
|
||
|
|
- **A**: Use Unicode character sets (e.g., `AL32UTF8`) for the target database if possible, and ensure `NLS_LANG` is correctly set to handle these characters.
|
||
|
|
|
||
|
|
---
|
||
|
|
|
||
|
|
### **Summary of Steps**
|
||
|
|
|
||
|
|
1. **Determine Character Sets**:
|
||
|
|
|
||
|
|
- Source and target database character sets.
|
||
|
|
|
||
|
|
2. **Set NLS_LANG for Export**:
|
||
|
|
|
||
|
|
- Match to source database character set.
|
||
|
|
|
||
|
|
3. **Export Data**:
|
||
|
|
|
||
|
|
- Use `exp` utility.
|
||
|
|
|
||
|
|
4. **Set NLS_LANG for Import**:
|
||
|
|
|
||
|
|
- Match to target database character set.
|
||
|
|
|
||
|
|
5. **Import Data**:
|
||
|
|
|
||
|
|
- Use `imp` utility.
|
||
|
|
|
||
|
|
6. **Verify Data Integrity**:
|
||
|
|
|
||
|
|
- Check for character corruption or data loss.
|
||
|
|
|
||
|
|
---
|
||
|
|
|
||
|
|
### **Example Commands**
|
||
|
|
|
||
|
|
#### **Export Command**
|
||
|
|
|
||
|
|
- **Windows**:
|
||
|
|
|
||
|
|
```cmd
|
||
|
|
SET NLS_LANG=AMERICAN_AMERICA.WE8MSWIN1252
|
||
|
|
exp userid=source_user/password@source_db file=export.dmp log=export.log
|
||
|
|
```
|
||
|
|
|
||
|
|
- **Unix/Linux**:
|
||
|
|
|
||
|
|
```bash
|
||
|
|
export NLS_LANG=AMERICAN_AMERICA.WE8MSWIN1252
|
||
|
|
exp userid=source_user/password@source_db file=export.dmp log=export.log
|
||
|
|
```
|
||
|
|
|
||
|
|
#### **Import Command**
|
||
|
|
|
||
|
|
- **Windows**:
|
||
|
|
|
||
|
|
```cmd
|
||
|
|
SET NLS_LANG=AMERICAN_AMERICA.AL32UTF8
|
||
|
|
imp userid=target_user/password@target_db file=export.dmp log=import.log
|
||
|
|
```
|
||
|
|
|
||
|
|
- **Unix/Linux**:
|
||
|
|
|
||
|
|
```bash
|
||
|
|
export NLS_LANG=AMERICAN_AMERICA.AL32UTF8
|
||
|
|
imp userid=target_user/password@target_db file=export.dmp log=import.log
|
||
|
|
```
|
||
|
|
|
||
|
|
---
|
||
|
|
|
||
|
|
### **Final Tips**
|
||
|
|
|
||
|
|
- **Backup**: Always backup your databases before performing import/export operations.
|
||
|
|
- **Documentation**: Keep records of your `NLS_LANG` settings and character sets for future reference.
|
||
|
|
- **Consult Oracle Documentation**: For complex scenarios, refer to Oracle's Globalization Support Guide.
|
||
|
|
|
||
|
|
---
|
||
|
|
|
||
|
|
By carefully setting the `NLS_LANG` environment variable to match the database character sets during export and import, you can safely transfer data between databases with different NLS_LANG settings. This approach ensures proper character set conversion and maintains data integrity across different environments.
|
||
|
|
|
||
|
|
If you have further questions or encounter specific issues during the process, feel free to ask for more assistance!
|