Conversation
ghost
self-assigned this
|
Note Reviews pausedIt looks like this branch is under active development. To avoid overwhelming you with review comments due to an influx of new commits, CodeRabbit has automatically paused this review. You can configure this behavior by changing the Use the following commands to manage reviews:
Use the checkboxes below for quick actions:
📝 WalkthroughWalkthroughReplaces the previous secondary-index how-to with two new documentation files: a conceptual overview and a practical how-to. All edits are documentation-only; no code, schema, or public API changes. Changes
Estimated code review effort🎯 2 (Simple) | ⏱️ ~10 minutes Suggested reviewers
🚥 Pre-merge checks | ✅ 3✅ Passed checks (3 passed)
✏️ Tip: You can configure your own custom pre-merge checks in the settings. Tip Try Coding Plans. Let us write the prompt for your AI agent so you can ship faster (with fewer bugs). Thanks for using CodeRabbit! It's free for OSS, and your support helps us grow. If you like it, consider giving us a shout-out. Comment |
![coderabbitai[bot]](https://avatars.githubusercontent.com/in/347564?s=60&v=4){kind=small}
There was a problem hiding this comment.
Actionable comments posted: 3
🤖 Fix all issues with AI agents
In `@docs/defradb/How-to` Guides/seconday-index-how-to.md:
- Around line 1-7: Rename the file named "seconday-index-how-to.md" to
"secondary-index-how-to.md" (fix the missing 'r'), and update any internal
references or links that point to "seconday-index-how-to.md" so they use
"secondary-index-how-to.md" instead; ensure the frontmatter and sidebar_label
remain unchanged and verify that any sidebar or table-of-contents entries
referencing the old filename are updated to the new filename.
- Line 25: The markdown link label "[secondary index concepts]" points to the
wrong path; update the link target from
"/defradb/next/Concepts/secondary-indexes" to the correct file name
"/defradb/next/Concepts/secondary-index.md" so the link resolves without a 404.
- Line 266: Update the link in the "Test query performance" bullet that
currently points to "/defradb/next/How-to%20Guides/explain-systems-how-to" so it
uses the same relative pattern as other How-to guides; replace that absolute
path with "./explain-systems-how-to.md" in the line containing "**Test query
performance**" to match existing guide linking conventions.
🧹 Nitpick comments (2)
docs/defradb/Concepts/secondary-index.md (2)
35-46: Consider adding language specifiers to code blocks for better rendering.The conceptual examples showing query execution would benefit from language specifiers (e.g.,
textorplaintext) to improve syntax highlighting and rendering consistency.💡 Suggested formatting improvement
-``` +```text Query: Find users with age = 30 Process: Scan all user documents → Check each age field → Return matches Cost: O(n) where n = total documents-
+text
Query: Find users with age = 30
Process: Look up "30" in age index → Return matching document IDs
Cost: O(log n) for lookup + O(m) for retrieval where m = matching documents
55-60: Add language specifiers to index structure examples.The code blocks showing index structures and formats lack language specifiers. Adding
textorplaintextidentifiers would improve rendering and satisfy linting rules.💡 Example fix for one block
-``` +```text Index entries: "Alice" → [doc_id_1] "Bob" → [doc_id_2, doc_id_3] "Charlie" → [doc_id_4]Apply similar changes to blocks at lines 94-98, 200-204, 213-222, and 272-280. </details> Also applies to: 94-98, 200-204, 213-222, 272-280 </blockquote></details> </blockquote></details> <details> <summary>📜 Review details</summary> **Configuration used**: Organization UI **Review profile**: CHILL **Plan**: Pro <details> <summary>📥 Commits</summary> Reviewing files that changed from the base of the PR and between 2b21cce44e2ae40d774891e99079c779873d6e3d and c5cd21b8d82847b8188587f2f25cb7306726d188. </details> <details> <summary>📒 Files selected for processing (3)</summary> * `docs/defradb/Concepts/secondary-index.md` * `docs/defradb/How-to Guides/secondary-index.md` * `docs/defradb/How-to Guides/seconday-index-how-to.md` </details> <details> <summary>💤 Files with no reviewable changes (1)</summary> * docs/defradb/How-to Guides/secondary-index.md </details> <details> <summary>🧰 Additional context used</summary> <details> <summary>🪛 markdownlint-cli2 (0.18.1)</summary> <details> <summary>docs/defradb/Concepts/secondary-index.md</summary> 12-12: Emphasis used instead of a heading (MD036, no-emphasis-as-heading) --- 35-35: Fenced code blocks should have a language specified (MD040, fenced-code-language) --- 42-42: Fenced code blocks should have a language specified (MD040, fenced-code-language) --- 55-55: Fenced code blocks should have a language specified (MD040, fenced-code-language) --- 94-94: Fenced code blocks should have a language specified (MD040, fenced-code-language) --- 200-200: Fenced code blocks should have a language specified (MD040, fenced-code-language) --- 213-213: Fenced code blocks should have a language specified (MD040, fenced-code-language) --- 219-219: Fenced code blocks should have a language specified (MD040, fenced-code-language) --- 272-272: Fenced code blocks should have a language specified (MD040, fenced-code-language) --- 277-277: Fenced code blocks should have a language specified (MD040, fenced-code-language) </details> </details> </details> <sub>✏️ Tip: You can disable this entire section by setting `review_details` to `false` in your review settings.</sub> </details> <!-- This is an auto-generated comment by CodeRabbit for review status -->
|
|
||
| - DefraDB installed and running | ||
| - A defined schema for your collections | ||
| - Understanding of [secondary index concepts](/defradb/next/Concepts/secondary-indexes) |
There was a problem hiding this comment.
Critical: Broken link to concepts page.
The link path uses secondary-indexes (plural) but the actual file is secondary-index.md (singular). This will result in a 404 error.
🔗 Proposed fix
-- Understanding of [secondary index concepts](/defradb/next/Concepts/secondary-indexes)
+- Understanding of [secondary index concepts](/defradb/next/Concepts/secondary-index)
📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| - Understanding of [secondary index concepts](/defradb/next/Concepts/secondary-indexes) | |
| - Understanding of [secondary index concepts](/defradb/next/Concepts/secondary-index) |
🤖 Prompt for AI Agents
In `@docs/defradb/How-to` Guides/seconday-index-how-to.md at line 25, The markdown
link label "[secondary index concepts]" points to the wrong path; update the
link target from "/defradb/next/Concepts/secondary-indexes" to the correct file
name "/defradb/next/Concepts/secondary-index.md" so the link resolves without a
404.
Deploying docs-source-network with
|
| Latest commit: |
07faa41
|
| Status: | ✅ Deploy successful! |
| Preview URL: | https://ee6b35c8.docs-source-network.pages.dev |
| Branch Preview URL: | https://seconday-index-update.docs-source-network.pages.dev |
islamaliev
left a comment
islamaliev
left a comment
There was a problem hiding this comment.
Overall documentation gives valuable information, but it looks mostly AI-generated and some information is incorrect.
Also I don't see any mention or arrays with indexes applied to them.
Also I feel like we can talk more about indexes on relations. Btw, 1-to-1 2-sided relations (in the example we have only 1-sided relation) are automatically constrained by a unique index to enforce 1-to-1 invariant.
Also I think we can add a warning about indexing arrays and json as this might fill up the storage pretty quickly with large data.
I didn't read "How-to Guide". First we need to adjust this one.
| ``` | ||
| Query: Find users with age = 30 | ||
| Process: Look up "30" in age index → Return matching document IDs | ||
| Cost: O(log n) for lookup + O(m) for retrieval where m = matching documents |
There was a problem hiding this comment.
todo: the cost for look up (like _eq matcher) is not O(log n), it's O(1)
| DefraDB stores indexes as sorted key-value pairs where: | ||
| - **Key**: The indexed field value(s) | ||
| - **Value**: Document identifier (_key) |
There was a problem hiding this comment.
todo: this is true only for unique indexes:
/col_id/ind_id/field_values -> _docID
For regular indexes it's
/col_id/ind_id/field_values/_docID -> {}
pointing to empty value
| "Charlie" → [doc_id_4] | ||
| ``` | ||
|
|
||
| When you query for `name = "Bob"`, DefraDB looks up "Bob" in the index and immediately retrieves `doc_id_2` and `doc_id_3`. |
There was a problem hiding this comment.
todo: it doesn't retrieve both at once. Depending on the condition it can retrieve only one of them, for example if limit: 1.
It will fetch them one-by-one
| ``` | ||
| ("published", "2024-01-15") → [doc_id_1] | ||
| ("published", "2024-01-16") → [doc_id_2, doc_id_3] | ||
| ("draft", "2024-01-15") → [doc_id_4] |
There was a problem hiding this comment.
todo: it's incorrect.
published/2024-01-15/doc_id_1 → {}
published/2024-01-16/doc_id_2 → {}
published/2024-01-16/doc_id_3 → {}
draft/2024-01-15/doc_id_4 → {}
Here I'm omitting col_id and index_id that are always prefixed.
ghost
force-pushed
the
Seconday-index-update
branch
from
c5cd21b to
899b3eb
Compare
ghost
force-pushed
the
Seconday-index-update
branch
from
899b3eb to
8a2549b
Compare
There was a problem hiding this comment.
Actionable comments posted: 1
📜 Review details
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (3)
docs/defradb/Concepts/secondary-index.mddocs/defradb/How-to Guides/secondary-index-how-to.mddocs/defradb/How-to Guides/secondary-index.md
💤 Files with no reviewable changes (1)
- docs/defradb/How-to Guides/secondary-index.md
🧰 Additional context used
📓 Path-based instructions (1)
**/*.{md,mdx}
📄 CodeRabbit inference engine (CLAUDE.md)
Documentation files support both
.md(CommonMark) and.mdx(MDX) formats
Files:
docs/defradb/How-to Guides/secondary-index-how-to.mddocs/defradb/Concepts/secondary-index.md
🪛 markdownlint-cli2 (0.21.0)
docs/defradb/Concepts/secondary-index.md
[warning] 35-35: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
[warning] 42-42: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
[warning] 55-55: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
[warning] 94-94: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
[warning] 200-200: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
[warning] 213-213: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
[warning] 219-219: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
[warning] 272-272: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
[warning] 277-277: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)
- GitHub Check: Cloudflare Pages
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@docs/defradb/Concepts/secondary-index.md`:
- Around line 35-39: The fenced code blocks in the Concept: secondary-index doc
lack language identifiers (markdownlint MD040); update each shown
triple-backtick block by adding an appropriate language tag (e.g., ```text or
```json) so they become fenced with a language: for example the block starting
with "Query: Find users with age = 30", the block starting "Process: Look up
\"30\" in age index → Return matching document IDs", and the other examples such
as the blocks beginning with "Index entries:", the tuple examples like
("published", "2024-01-15") → [doc_id_1], the nested path arrays like
["user","device","model","iPhone"], the doc_id mapping blocks, and the
collection/index path examples — prepend each opening ``` with a language token
(suggestion: text for plain examples, json for JSON-like snippets).
---
Duplicate comments:
In `@docs/defradb/How-to` Guides/secondary-index-how-to.md:
- Line 266: Replace the absolute link in the "Test query performance" bullet
("Use the [explain
systems](/defradb/next/How-to%20Guides/explain-systems-how-to) to analyze query
execution") with a relative link to the local How‑to page (e.g., change to "Use
the [explain systems](./explain-systems-how-to.md) to analyze query execution")
so it matches other How‑to guides' relative link format; update the link target
string in that exact bullet text.
There was a problem hiding this comment.
Actionable comments posted: 1
📜 Review details
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (2)
docs/defradb/Concepts/secondary-index.mddocs/defradb/How-to Guides/secondary-index-how-to.md
🚧 Files skipped from review as they are similar to previous changes (1)
- docs/defradb/Concepts/secondary-index.md
🧰 Additional context used
📓 Path-based instructions (1)
**/*.{md,mdx}
📄 CodeRabbit inference engine (CLAUDE.md)
Documentation files support both
.md(CommonMark) and.mdx(MDX) formats
Files:
docs/defradb/How-to Guides/secondary-index-how-to.md
🪛 LanguageTool
docs/defradb/How-to Guides/secondary-index-how-to.md
[style] ~15-~15: This phrase might be redundant. Consider either removing or replacing the adjective ‘additional’.
Context: ... and use unique indexes sparingly (they add an additional read operation on every write). Plan indexes...
(ADD_AN_ADDITIONAL)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)
- GitHub Check: Cloudflare Pages
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@docs/defradb/How-to` Guides/secondary-index-how-to.md:
- Line 15: Replace the phrase "an additional read operation" in the sentence
that warns about unique indexes with a tighter wording such as "an extra read
operation" or simply "a read operation"; locate the sentence containing "unique
indexes sparingly (they add an additional read operation on every write)" and
update that parenthetical to use the preferred phrasing.
---
Duplicate comments:
In `@docs/defradb/How-to` Guides/secondary-index-how-to.md:
- Line 286: In the sentence that links to the explain systems guide, replace the
absolute URL "/defradb/next/How-to%20Guides/explain-systems-how-to" with a
relative link matching other How-to guides (e.g., ./explain-systems-how-to.md);
update the link text near "explain systems" in secondary-index-how-to.md so it
uses the relative path and the same markdown link style as sibling files.
There was a problem hiding this comment.
Actionable comments posted: 3
📜 Review details
Configuration used: Organization UI
Review profile: CHILL
Plan: Pro
📒 Files selected for processing (1)
docs/defradb/Concepts/secondary-index.md
🧰 Additional context used
📓 Path-based instructions (1)
**/*.{md,mdx}
📄 CodeRabbit inference engine (CLAUDE.md)
Documentation files support both
.md(CommonMark) and.mdx(MDX) formats
Files:
docs/defradb/Concepts/secondary-index.md
🪛 markdownlint-cli2 (0.21.0)
docs/defradb/Concepts/secondary-index.md
[warning] 37-37: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
[warning] 45-45: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
[warning] 55-55: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
[warning] 61-61: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
[warning] 67-67: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
[warning] 108-108: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
[warning] 228-228: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
[warning] 242-242: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
[warning] 248-248: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
[warning] 291-291: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
[warning] 297-297: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
⏰ Context from checks skipped due to timeout of 90000ms. You can increase the timeout in your CodeRabbit configuration to a maximum of 15 minutes (900000ms). (1)
- GitHub Check: Cloudflare Pages
🤖 Prompt for all review comments with AI agents
Verify each finding against the current code and only fix it if needed.
Inline comments:
In `@docs/defradb/Concepts/secondary-index.md`:
- Line 320: The single-field example showing publishedAt: DateTime
`@index`(direction: DESC) is incorrectly placed under the "composite index"
direction subsection; replace that single-field snippet with a composite-index
example that matches the heading — use an `@index`(includes: [...]) example
showing per-field directions (e.g., {field: "status", direction: ASC}, {field:
"publishedAt", direction: DESC}) and move or re-add the single-field publishedAt
example to the earlier single-field discussion if you want to keep it as an
illustration; ensure the composite `@index`(includes: ...) snippet replaces the
current single-field code block so the prose and example align.
- Line 198: Replace the non-standard phrase "1-to-2-sided relations" with a
clearer term such as "one-to-one two-sided (bidirectional) relations" and update
the note so it reads along the lines of: "Note: one-to-one two-sided
(bidirectional) relations are automatically constrained by a unique index to
enforce the one-to-one invariant." Ensure the replacement appears wherever the
original phrase appears in the Concepts/secondary-index.md content and preserve
the intent about automatic unique-index enforcement.
- Line 45: Update the documented index lookup complexity in
docs/defradb/Concepts/secondary-index.md: replace the incorrect "O(1)"
complexity with "O(log n)" in the two locations called out (the text around line
48 and the paragraph around line 308) and mention that this reflects BadgerDB's
LSM-tree behavior (memtable + multiple SSTable levels) so point lookups are
O(log n) in the worst case.
---
Duplicate comments:
In `@docs/defradb/Concepts/secondary-index.md`:
- Around line 37-41: The fenced code blocks in
docs/defradb/Concepts/secondary-index.md are missing language identifiers
(MD040); update each bare triple-backtick fence to include a language tag (use
"text") for the examples such as the "Query: Find users with age = 30" block,
the index lookup examples (e.g. lines containing
"/col_id/ind_id/field_values/_docID → {}",
"/1/1/user.device.model/iPhone/doc_id_1 → {}", and the
"<collection_id>/<index_id>/<json_path>/<json_value>/<doc_id>" sample), and
apply the same change to all other reported ranges (45-49, 55-57, 61-63, 67-73,
108-113, 228-232, 242-244, 248-251, 291-293, 297-300) so every ``` becomes
```text.
|
|
||
| **With an index on age:** | ||
|
|
||
| ``` |
There was a problem hiding this comment.
🧩 Analysis chain
🏁 Script executed:
cat -n docs/defradb/Concepts/secondary-index.md | sed -n '40,55p'
Repository: sourcenetwork/docs.source.network
Length of output: 598
🏁 Script executed:
cat -n docs/defradb/Concepts/secondary-index.md | sed -n '300,315p'
Repository: sourcenetwork/docs.source.network
Length of output: 1059
🏁 Script executed:
wc -l docs/defradb/Concepts/secondary-index.md
Repository: sourcenetwork/docs.source.network
Length of output: 121
🏁 Script executed:
cd docs && find . -type f \( -name "*.md" -o -name "*.mdx" \) -exec grep -l "BadgerDB\|storage engine\|LSM" {} \;
Repository: sourcenetwork/docs.source.network
Length of output: 173
🏁 Script executed:
# Search for DefraDB architecture documentation
find docs -type f \( -name "*.md" -o -name "*.mdx" \) | xargs grep -i "backend\|storage" | head -20
Repository: sourcenetwork/docs.source.network
Length of output: 5221
🏁 Script executed:
# Look for any mentions of indexing complexity or performance
cat -n docs/defradb/Concepts/secondary-index.md | grep -i "o(1)\|o(log\|o(n)\|complexity\|time"
Repository: sourcenetwork/docs.source.network
Length of output: 752
🏁 Script executed:
# Search for BadgerDB mentions
find docs -type f \( -name "*.md" -o -name "*.mdx" \) | xargs grep -i "badger" 2>/dev/null
Repository: sourcenetwork/docs.source.network
Length of output: 664
🏁 Script executed:
# Check deployment.md for backend info
cat -n docs/defradb/How-to\ Guides/deployment.md | head -50
Repository: sourcenetwork/docs.source.network
Length of output: 3319
🏁 Script executed:
# Search for DefraDB core documentation about storage
find docs -type f -name "*.md" | xargs grep -i "storage engine\|database backend\|merkle" 2>/dev/null | head -15
Repository: sourcenetwork/docs.source.network
Length of output: 3389
🏁 Script executed:
# Look at the main DefraDB documentation structure
ls -la docs/defradb/
Repository: sourcenetwork/docs.source.network
Length of output: 572
🏁 Script executed:
cat -n docs/defradb/Concepts/secondary-index.md | sed -n '50,100p'
Repository: sourcenetwork/docs.source.network
Length of output: 1694
🏁 Script executed:
# Check if there's more context about index implementation
cat -n docs/defradb/Concepts/secondary-index.md | sed -n '1,60p'
Repository: sourcenetwork/docs.source.network
Length of output: 2792
🌐 Web query:
BadgerDB LSM tree lookup complexity point query performance
💡 Result:
Web search failed: Server error: no LLM provider could handle the message
Update index lookup complexity from O(1) to O(log n) to reflect BadgerDB's LSM-tree implementation.
DefraDB uses BadgerDB (an LSM-tree-based key-value store) as its default storage backend. Point lookups in LSM trees have worst-case complexity of O(log n) due to checking the memtable and potentially multiple SSTable levels, not O(1). Lines 48 and 308 should reflect this accurate complexity.
Changes needed
Line 48:
-Cost: O(1) for lookup + O(m) for retrieval where m = matching documents
+Cost: O(log n) for lookup + O(m) for retrieval where m = matching documents
Line 308:
-On reads, an `_eq` filter on an indexed field is O(1) for the lookup, plus O(m) to retrieve the m matching documents.
+On reads, an `_eq` filter on an indexed field is O(log n) for the lookup, plus O(m) to retrieve the m matching documents.
🧰 Tools
🪛 markdownlint-cli2 (0.21.0)
[warning] 45-45: Fenced code blocks should have a language specified
(MD040, fenced-code-language)
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@docs/defradb/Concepts/secondary-index.md` at line 45, Update the documented
index lookup complexity in docs/defradb/Concepts/secondary-index.md: replace the
incorrect "O(1)" complexity with "O(log n)" in the two locations called out (the
text around line 48 and the paragraph around line 308) and mention that this
reflects BadgerDB's LSM-tree behavior (memtable + multiple SSTable levels) so
point lookups are O(log n) in the worst case.
|
|
||
| Without the unique constraint, the relationship defaults to one-to-many (multiple Users could reference the same Address). The unique index ensures exactly one User per Address. | ||
|
|
||
| Note: 1-to-2-sided relations are automatically constrained by a unique index to enforce the 1-to-1 invariant. |
There was a problem hiding this comment.
Clarify the non-standard "1-to-2-sided" terminology.
"1-to-2-sided relations" is not a recognised cardinality term and is likely confusing to readers. Based on context (enforcing a 1-to-1 invariant), this probably means "one-to-one two-sided (bidirectional)" relations.
🔧 Proposed fix
-Note: 1-to-2-sided relations are automatically constrained by a unique index to enforce the 1-to-1 invariant.
+Note: One-to-one (bidirectional) relations are automatically constrained by a unique index to enforce the 1-to-1 invariant.
📝 Committable suggestion
‼️ IMPORTANT
Carefully review the code before committing. Ensure that it accurately replaces the highlighted code, contains no missing lines, and has no issues with indentation. Thoroughly test & benchmark the code to ensure it meets the requirements.
| Note: 1-to-2-sided relations are automatically constrained by a unique index to enforce the 1-to-1 invariant. | |
| Note: One-to-one (bidirectional) relations are automatically constrained by a unique index to enforce the 1-to-1 invariant. |
🤖 Prompt for AI Agents
Verify each finding against the current code and only fix it if needed.
In `@docs/defradb/Concepts/secondary-index.md` at line 198, Replace the
non-standard phrase "1-to-2-sided relations" with a clearer term such as
"one-to-one two-sided (bidirectional) relations" and update the note so it reads
along the lines of: "Note: one-to-one two-sided (bidirectional) relations are
automatically constrained by a unique index to enforce the one-to-one
invariant." Ensure the replacement appears wherever the original phrase appears
in the Concepts/secondary-index.md content and preserve the intent about
automatic unique-index enforcement.
|
|
||
| ## Direction and ordering | ||
|
|
||
| Index direction (ASC or DESC) plays a significant role primarily for **composite indexes**. For single-field indexes, the index fetcher can traverse entries in reverse order just as efficiently as the default order, so direction has minimal practical impact there. |
There was a problem hiding this comment.
Single-field direction example is misplaced under the "composite index" direction sub-section.
Line 322 says "For composite indexes, specifying direction can matter:" but the code block that immediately follows (Lines 324–328) demonstrates a single-field index (publishedAt: DateTime @index(direction: DESC)). The composite example doesn't appear until Lines 332–337. This contradicts the preceding prose (Line 320) which explicitly states direction has minimal impact on single-field indexes.
Either move the single-field snippet to illustrate the single-field case earlier, or replace it with a composite index example that matches the section heading.
🔧 Proposed fix — replace misplaced single-field example with the composite one
For composite indexes, specifying direction can matter:
-```graphql
-type Article {
- publishedAt: DateTime `@index`(direction: DESC)
-}
-```
-
-Each field in a composite index can have its own direction:
-
```graphql
`@index`(includes: [
{field: "status", direction: ASC},
{field: "publishedAt", direction: DESC}
])
</details>
Also applies to: 322-322, 324-324, 328-328
<details>
<summary>🤖 Prompt for AI Agents</summary>
Verify each finding against the current code and only fix it if needed.
In @docs/defradb/Concepts/secondary-index.md at line 320, The single-field
example showing publishedAt: DateTime @index(direction: DESC) is incorrectly
placed under the "composite index" direction subsection; replace that
single-field snippet with a composite-index example that matches the heading —
use an @index(includes: [...]) example showing per-field directions (e.g.,
{field: "status", direction: ASC}, {field: "publishedAt", direction: DESC}) and
move or re-add the single-field publishedAt example to the earlier single-field
discussion if you want to keep it as an illustration; ensure the composite
@index(includes: ...) snippet replaces the current single-field code block so
the prose and example align.
</details>
<!-- fingerprinting:phantom:triton:churro -->
<!-- This is an auto-generated comment by CodeRabbit -->
| /1/1/user.device.model/Android/doc_id_2 → {} | ||
| ``` | ||
|
|
||
| When you query for a specific path and value, DefraDB directly looks it up in the inverted index and retrieves all matching documents. For more on inverted indexes, see the [CockroachDB RFC on inverted indexes](https://github.com/cockroachdb/cockroach/blob/master/docs/RFCS/20171020_inverted_indexes.md). |
There was a problem hiding this comment.
I'm not sure if we should reference CocroachDB in our documentation. @jsimnz what do you think?
| } | ||
| ``` | ||
|
|
||
| This ensures no two Users can reference the same Address document. Note that 1-to-2-sided relations are automatically constrained by a unique index to enforce the 1-to-1 invariant. |
There was a problem hiding this comment.
todo: I'm not familiar with the term "1-to-2-sided relations". There is "1-to-1" and "1-to-many"
2 participants
Uh oh!
There was an error while loading. Please reload this page.