This is an automated email from the ASF dual-hosted git repository.
morningman pushed a commit to branch orc
in repository https://gitbox.apache.org/repos/asf/doris-thirdparty.git
The following commit(s) were added to refs/heads/orc by this push:
new c09e9df6a04 [fix] Fix set reader category in late materialization.
(#257)
c09e9df6a04 is described below
commit c09e9df6a04e5c48ad6c49a339d589d8c1898615
Author: Qi Chen <che...@selectdb.com>
AuthorDate: Wed Nov 27 22:05:51 2024 +0800
[fix] Fix set reader category in late materialization. (#257)
---
c++/src/Reader.cc | 47 +++++++++++++++++++++++++++++++++++++++++++++++
1 file changed, 47 insertions(+)
diff --git a/c++/src/Reader.cc b/c++/src/Reader.cc
index eddeeee0b05..d721867622e 100644
--- a/c++/src/Reader.cc
+++ b/c++/src/Reader.cc
@@ -340,6 +340,53 @@ namespace orc {
throw ParseError("Invalid column selected " + colName);
}
}
+
+ std::unordered_set<int> filterColIds;
+ if (!filterCols.empty()) {
+ for (const auto& colName : filterCols) {
+ auto iter = nameTypeMap.find(colName);
+ if (iter == nameTypeMap.end()) {
+ throw ParseError("Invalid column selected " + colName);
+ }
+
+ Type* type = iter->second;
+
+ // Process current node and all its parent nodes
+ // Set FILTER_CHILD for leaf nodes and FILTER_PARENT for non-leaf nodes
+ Type* current = type;
+ while (current != nullptr) {
+ if (current->getSubtypeCount() == 0) {
+ current->setReaderCategory(ReaderCategory::FILTER_CHILD);
+ } else {
+ current->setReaderCategory(ReaderCategory::FILTER_PARENT);
+ }
+ filterColIds.emplace(current->getColumnId());
+ current = current->getParent();
+ }
+
+ // Process all child nodes of the current node
+ // For child nodes: set FILTER_PARENT if it's a leaf, FILTER_CHILD if
it has children
+ std::function<void(Type*)> processChildren = [&processChildren](Type*
node) {
+ if (node == nullptr) return;
+
+ // Iterate through all child nodes
+ for (int i = 0; i < node->getSubtypeCount(); ++i) {
+ Type* child = node->getSubtype(i);
+ if (child->getSubtypeCount() == 0) {
+ // Leaf node (no children)
+ child->setReaderCategory(ReaderCategory::FILTER_PARENT);
+ } else {
+ // Non-leaf node (has children)
+ child->setReaderCategory(ReaderCategory::FILTER_CHILD);
+ // Recursively process its children
+ processChildren(child);
+ }
+ }
+ };
+
+ processChildren(type);
+ }
+
startReadPhase = ReadPhase::LEADERS;
readerContext = std::unique_ptr<ReaderContext>(new ReaderContext());
readerContext->setFilterCallback(std::move(filterColIds), filter);
---------------------------------------------------------------------
To unsubscribe, e-mail: commits-unsubscr...@doris.apache.org
For additional commands, e-mail: commits-h...@doris.apache.org
