TiDB substring incorrectly treats start argument as signed integer
- Dominant language
- Go
- Stars
- 40.5k
- Forks
- 6.2k
- PR merge metrics
- PR metrics pending
Description
## Bug Report
Please answer these questions before submitting your issue. Thanks!
### 1. Minimal reproduce step (Required)
prepare data
```sql
drop table if exists t0;
CREATE TABLE t0(c0 CHAR PRIMARY KEY);
REPLACE INTO t0 VALUES ('1'), ('1');
ALTER TABLE t0 SET TIFLASH REPLICA 1;
```
execute query
```sql
SELECT /*+ read_from_storage(tikv[t0]) */ * FROM t0 WHERE SUBSTRING(1395944821, (~ (t0.c0)));
SELECT /*+ read_from_storage(tiflash[t0]) */ * FROM t0 WHERE SUBSTRING(1395944821, (~ (t0.c0)));
```
### 2. What did you expect to see? (Required)
The results of the two queries are equal.
### 3. What did you see instead (Required)
Query1:

Query2: empty
### 4. What is your TiDB version? (Required)
Release Version: v8.4.0
Edition: Community
Git Commit Hash: 1a9f0fa25580907d3266dc3edf000ba24379a600
Git Branch: HEAD
UTC Build Time: 2024-11-07 15:18:43
GoVersion: go1.23.2
Race Enabled: false
Check Table Before Drop: false
Store: tikv
Contributor guide
Research direction
Start by running the provided SUBSTRING queries against the TiKV and TiFlash storage paths on TiDB v8.4.0, then compare how the signed start argument is evaluated for each path. Done means both queries return equal results for the reproduction case; the issue does not name source files or tests to guide implementation.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- go, sql
- Domain
- databases, distributed-systems
- Issue type
- Bug
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100