Skip to content

Commit 0935e67

Browse files
committed
boilerplate
Signed-off-by: Adam Gutglick <adam@spiraldb.com>
1 parent b7b8633 commit 0935e67

1 file changed

Lines changed: 71 additions & 0 deletions

File tree

rfcs/0063-map-type.md

Lines changed: 71 additions & 0 deletions
Original file line numberDiff line numberDiff line change
@@ -0,0 +1,71 @@
1+
- Start Date: 2026-07-16
2+
- Authors: @AdamGS
3+
- RFC PR: https://github.com/vortex-data/rfcs/pull/63
4+
5+
# Map Type
6+
7+
## Summary
8+
9+
One paragraph explanation of the proposed map type.
10+
11+
## Motivation
12+
13+
What problem does this solve? Include concrete use cases where possible.
14+
15+
- What specific use cases does this enable or improve?
16+
- What workflows or operations are painful, slow, or impossible today?
17+
18+
## Design
19+
20+
Describe the proposed design in enough detail that someone familiar with Vortex could implement it.
21+
This should cover:
22+
23+
- New or modified APIs, traits, or vtable entries.
24+
- How this interacts with existing components (encodings, layouts, scan, file format, etc.).
25+
- Key implementation details and corner cases.
26+
- Why is this the best approach in the space of possible designs?
27+
- Which crates are affected and how the dependency graph changes, if at all.
28+
29+
Use code examples and diagrams where they might help.
30+
31+
## Compatibility
32+
33+
- Does this change the file format or wire format? Is it backward/forward compatible?
34+
- Does this break any public APIs? If so, what is the migration path?
35+
- Are there performance implications?
36+
37+
If there are no compatibility concerns, briefly state why.
38+
39+
## Drawbacks
40+
41+
- Why should we _not_ do this?
42+
- What is the maintenance cost of this change?
43+
- Does this add complexity that could be avoided?
44+
45+
## Alternatives
46+
47+
- What other designs were considered and why were they rejected?
48+
- What is the cost of **not** doing this?
49+
- Is there a simpler approach that gets us most of the way there?
50+
51+
## Prior Art
52+
53+
How have other systems solved this or similar problems? Consider:
54+
55+
- Other columnar formats (Parquet, Arrow, etc.).
56+
- Database internals (DuckDB, DataFusion, Velox, etc.).
57+
- Relevant academic papers or blog posts.
58+
59+
This section helps frame the design in a broader context.
60+
If there is no relevant prior art, that is fine.
61+
62+
## Unresolved Questions
63+
64+
- What parts of the design need to be resolved during the RFC process?
65+
- What is explicitly out of scope for this RFC?
66+
- Are there open questions that can be deferred to implementation?
67+
68+
## Future Possibilities
69+
70+
What natural extensions or follow-on work does this enable? This is a good place to note related
71+
ideas that are out of scope for this RFC but worth capturing.

0 commit comments

Comments
 (0)