[TiDB Community Survey] Long-Standing Search Issue Optimization

[translator_bot]

Note:
This topic has been translated from a Chinese forum by GPT and might contain errors.

Original topic: 【TiDB 社区调研】吐槽已久的搜索问题优化

| username: TiDB社区小助手

---

Research Background

The “search issue” has long been a pain point for TiDBers, and the TiDB community has been continuously working to optimize this problem. Every optimization in the community is inseparable from the contributions and feedback of TiDBers. We believe that TiDBers who use the community forum daily have a deeper understanding of the “search issue.” This survey aims to invite TiDBers to discuss specific optimization suggestions: In what content searches are the results inaccurate? Every piece of feedback will be carefully recorded. The community’s search function is changing because of you.

Feedback Format

Please provide specific examples, including four elements:

1. Search keyword: XXX
2. Screenshot of the search results page
3. Expected content
4. Optimization suggestions

For example:

1. Search keyword: Chat Session 44

2. Screenshot of the search results page👇
   

3. Expected content: The first search result should be Chat Session 44

4. Optimization suggestions: Adjust the accuracy of search results, prioritizing related results instead of showing Session 22 first

Feedback Rewards

1. Provide feedback according to the format: Earn 100 experience points, 100 points
2. Participate in the lottery: TiDBers who provide feedback according to the format will be entered into a draw to win 2 TiDB thermos cups (Winter is coming, let the community thermos cup warm you up～)
   

Feedback Period

2022.11.11-2022.11.28

[translator_bot]

| username: 数据小黑 | Original post link

---

1. Search keyword: Activity

2. Screenshot of the search results page
   

   

3. Expected content
   

4. Optimization suggestions
   For some keywords, is it possible to prepare some specific search conditions? For example, when searching for activities, the intention is to query the list of posts initiated by the official, sorted in reverse chronological order. Actually, the current search has improved a lot compared to before. This small suggestion might not be very universal, just a humble opinion to spark some ideas.

[translator_bot]

| username: onlyacat | Original post link

---

Please optimize it well. I don’t even use the forum’s internal search anymore; I just use Google with a specified domain . The results are more accurate.

[translator_bot]

| username: caiyfc | Original post link

---

1. Search keyword: Statistics Information

2. Search page screenshot
   

   

3. Expected content: Expect the first search result to be Introduction to Statistics | PingCAP Docs

4. Optimization suggestion: Adjust the search result accuracy to prioritize displaying major title content, rather than the statistics information in TiSpark.

[translator_bot]

| username: xfworld | Original post link

---

1. Search keyword: Lock conflict

2. Screenshot of the search results page
   

   
   
   
   

3. Expected content
   It should be helpful, and when clicked, it should have relevant content. Instead, it directly shows a 404 error.

4. Optimization suggestions
   Ensure the links are real and valid, and avoid 404 errors…

[translator_bot]

| username: alfred | Original post link

---

1. Search keyword: LSM-Tree

2. Search page screenshot

3. Expected content
   Able to accurately provide content

4. Optimization suggestions
   Did not see the best solution on the homepage

[translator_bot]

| username: 半瓶醋仙 | Original post link

---

1. Search keyword: TiDBer
2. Search page screenshot:
   
   
3. Expected content: Add filtering and grading options to the search results, include time sorting, attribute content sorting, topic sorting, and custom sorting functions.
4. Optimization suggestion: Display the time taken for the search.

[translator_bot]

| username: coderv | Original post link

---

1. Search keyword: Environment

2. Search page screenshot
   

   

3. Expected content: Expect the search results to have distinctive titles

4. Optimization suggestion: The information displayed in the search results is too limited, sometimes it’s hard to distinguish the content of each page. It feels like it could be optimized.

[translator_bot]

| username: 张雨齐0720 | Original post link

---

1. Search keyword: HTAP solution
2. Search page screenshot
   
   
3. Expected content: The first search result should be an overall introduction to TiDB’s HTAP implementation capabilities.
4. Optimization suggestion: Adjust the search result accuracy to prioritize the implementation of HTAP solutions rather than directly displaying the introduction of TiDB.

[translator_bot]

| username: Jiawei | Original post link

---

1. Search keywords: TiDB lock
2. Page screenshot
   
   
3. Expected content
   I hope the official documentation can be prioritized over official blogs and columns, sorted from the most authoritative to the least.
4. Optimization suggestion: Prioritize displaying the most authoritative technical documents.

[translator_bot]

| username: Billmay表妹 | Original post link

---

This is a BUG, I will report it.

[translator_bot]

| username: Billmay表妹 | Original post link

---

Which one is the optimal solution? Is it because the content exists but cannot be found, or is it because it doesn’t exist at all?

[translator_bot]

| username: alfred | Original post link

---

This result is still pretty good.

[translator_bot]

| username: Kongdom | Original post link

---

1. Search keyword: request is too large
2. Search page screenshot👇
   
   
3. Expected content:
   The first search result should be TICDC报错ErrGCTTLExceeded，etcdserver: request is too large
4. Optimization suggestion: Adjust the search result accuracy to prioritize displaying related results, rather than showing “too large” first.

[translator_bot]

| username: 裤衩儿飞上天 | Original post link

---

1. Search keyword: database time zone difference of 8 hours

2. Screenshot of the search results page
   

   

3. Expected content
   Time zone support, time zone handling related classes should be among the top 5 search results

4. Optimization suggestions
   Adjust the accuracy of search results, prioritize displaying related results

[translator_bot]

| username: dockerfile | Original post link

---

Got it.

[translator_bot]

| username: TiDBer_jYQINSnf | Original post link

---

It is expected to add a feature to search for all answers by a specific user, with the ability to filter out all answers marked as solutions, and to add a time filter.
In other words, don’t just have a single search box for everything; provide advanced search options to narrow down the search scope with some clues.

Alternatively, provide a feature to continue searching within the search results, similar to grepping the content and then grepping again.

[translator_bot]

| username: Hacker007 | Original post link

---

1. Search keyword: XXX

2. Screenshot of the search results page
   

   

3. Expected content
   

   

4. Add search term highlighting, so I don’t have to find which result matches my query each time.

[translator_bot]

| username: Jiawei | Original post link

---

1. Search keywords
2. Search screenshot, click in and the version is missing
   
   
   
   
   
3. Expected content
   I hope that when I click in, I can see the pages corresponding to all versions. Sometimes, the search results do not match the version I am using. It would be convenient if I could quickly switch to my corresponding version’s documentation to see if there is a solution, instead of reading through an entire solution for version 6 when my version does not have that option.
4. Suggestion
   Display all mainstream versions.

[translator_bot]

| username: Billmay表妹 | Original post link

---

So you want this feature to have a version filter, right?