add scan edge/vertex in storage #1381
Conversation
critical27
added
ready-for-testing
critical27
requested review from
dangleptr,
jude-zhu,
laura-ding and
liuyu85cn
|
Jenkins go |
|
Unit testing passed. |
1 similar comment
|
Unit testing passed. |
|
Unit testing failed. |
|
Unit testing passed. |
|
Unit testing passed. |
|
Unit testing passed. |
|
Unit testing passed. |
|
Unit testing passed. |
src/kvstore/RocksEngine.h
Outdated
|
|
||
| ~RocksRangeWithPrefixIter() = default; | ||
|
|
||
| bool valid() const override { |
There was a problem hiding this comment.
Why override the method? It looks the same.
| } | ||
|
|
||
| EdgeType edgeType = NebulaKeyUtils::getEdgeType(key); | ||
| if (edgeType < 0) { |
|
|
||
| if (returnAllColumns_) { | ||
| // return all columns | ||
| edgeContexts_.emplace(edgeType, std::move(propContexts)); |
There was a problem hiding this comment.
Do we really need to set edgeContexts_ and edgeSchema_ when returning all columns
There was a problem hiding this comment.
edgeContexts_ and edgeSchema_ are used to store related edgeType, so we can check if we need to scan this edge
|
Unit testing passed. |
|
Unit testing passed. |
|
Unit testing passed. |
|
Unit testing passed. |
|
Unit testing passed. |
|
Unit testing passed. |
| @@ -185,6 +185,12 @@ class NebulaKeyUtils final { | |||
| return readInt<EdgeRanking>(rawKey.data() + offset, sizeof(EdgeRanking)); | |||
| } | |||
|
|
|||
| static int64_t getVersion(const folly::StringPiece& rawKey) { | |||
There was a problem hiding this comment.
Is it better to use a macro definition instead of int64_t? like VertexID, EdgeRanking
There was a problem hiding this comment.
We use int64_t as timestamp for now, maybe we can unify them with a macro later in another PR, WDYT?
There was a problem hiding this comment.
Sorry, I just found out that the macro has already existed.
Using SchemaVer
There was a problem hiding this comment.
I thought about it later, SchemaVer is used in tag or edge schema to indicate version. Should we define a DataVer to indicate the version of tag or edge data?
| std::string cursor = ""; | ||
| int32_t rowCount = 0, expectRowCount = 100; | ||
| checkResponse(partId, resp, cursor, rowCount, expectRowCount, 0, false, true); | ||
| EXPECT_EQ(rowCount, 100); |
There was a problem hiding this comment.
For edges, when no output column attributes are specified. What is output?
Does it output the following four fields：_src, _rank, _dst, _type?
There was a problem hiding this comment.
We only return the key, aka, src, type and dst (we don't return rank for now)
|
Unit testing passed. |
| auto f = processor->getFuture(); | ||
| processor->process(req); | ||
| auto resp = std::move(f).get(); | ||
| ++batchCount; |
There was a problem hiding this comment.
batchCount is not used.
There was a problem hiding this comment.
Never mind, we have count the total rows returned
* add scan edge/vertex, rebased * refactor scan edge/vertex * address @panda-sheep's comments Co-authored-by: dangleptr <37216992+dangleptr@users.noreply.github.com>
Co-authored-by: jimingquan <mingquan.ji@vesoft.com>
Support scan edge/vertex in storage, later for spark job.
We can retrieve data of a part limited by single vertexId and time range. We could enhance it by adding complicated filter and group of vertex later.