terrymanu commented on issue #34088:
URL:
https://github.com/apache/shardingsphere/issues/34088#issuecomment-3536718853
Analysis Conclusion: This is not a code defect in ShardingSphere, but rather
a misunderstanding of ShardingSphere's design philosophy.
## ShardingSphere Design Philosophy
ShardingSphere's core design goal is to enable users to use distributed
databases as if they were using a single database. Under this design philosophy:
1. Unified Table Space: One logical database should have a unified table
name space
2. Handling Same-Named Tables: Tables with the same name across different
data sources should be treated as sharded tables, not independent single tables
## Root Cause Analysis
The current configuration violates ShardingSphere's fundamental design
principles:
- Incorrectly configuring same-named tables from different data sources as
single tables
- Expecting to independently access same-named tables within the same
logical database
## Correct Solutions
Option 1: Configure as Sharded Tables (Recommended)
```yaml
spring:
shardingsphere:
rules:
sharding:
tables:
t_user: # Unified logical table name
actual-data-nodes: ds0.t_user, ds1.t_user
table-strategy:
standard:
sharding-column: user_id
algorithm-expression: t_user_${user_id % 2}
default-database-strategy:
standard:
sharding-column: user_id
algorithm-expression: ds${user_id % 2}
```
Option 2: Use Different Logical Table Names
If you truly need independent access for business reasons, use different
logical table names:
```yaml
spring:
shardingsphere:
rules:
sharding:
tables:
user_ds0: # User table in ds0
actual-data-nodes: ds0.user
user_ds1: # User table in ds1
actual-data-nodes: ds1.user
```
## Configuration Recommendations
Please choose the appropriate configuration based on your business
requirements:
1. If data needs horizontal sharding → Use sharded table configuration
2. If data is completely independent → Use different logical table names
3. If cross-datasource access is needed → Consider using different logical
databases
## Conclusion
ShardingSphere is designed to provide a transparent distributed database
experience, not to manage multiple independent databases. Same-named tables
within the same logical database should be managed uniformly through
sharding rules, not exist as independent single tables.
We recommend revisiting your data architecture design and choosing a
configuration approach that aligns with ShardingSphere's design philosophy.
--
This is an automated message from the Apache Git Service.
To respond to the message, please log on to GitHub and use the
URL above to go to the specific comment.
To unsubscribe, e-mail: [email protected]
For queries about this service, please contact Infrastructure at:
[email protected]
