This is an automated email from the ASF dual-hosted git repository. leirui pushed a commit to branch research/LTS-visualization in repository https://gitbox.apache.org/repos/asf/iotdb.git
commit 6cb1c18f279307e45d8f490b63e997839a194f6a Author: Lei Rui <[email protected]> AuthorDate: Sat Feb 3 16:39:35 2024 +0800 improve --- .../groupby/LocalGroupByExecutorTri_ILTS.java | 61 +++++++++++++++++++--- 1 file changed, 53 insertions(+), 8 deletions(-) diff --git a/server/src/main/java/org/apache/iotdb/db/query/dataset/groupby/LocalGroupByExecutorTri_ILTS.java b/server/src/main/java/org/apache/iotdb/db/query/dataset/groupby/LocalGroupByExecutorTri_ILTS.java index 25a9ccc6f71..82c1e06262b 100644 --- a/server/src/main/java/org/apache/iotdb/db/query/dataset/groupby/LocalGroupByExecutorTri_ILTS.java +++ b/server/src/main/java/org/apache/iotdb/db/query/dataset/groupby/LocalGroupByExecutorTri_ILTS.java @@ -384,10 +384,40 @@ public class LocalGroupByExecutorTri_ILTS implements GroupByExecutor { // 否则块完整落在桶内时，用凸包规则快速找到这个块中沿着lr连线法向量最高和最低的点，最后和全局当前最远结果点比较 // 也可以改成先不管是不是完整落在桶里，先找到最高低点，然后如果这两个点没有当前已知最远点远那就可以排除了， // 否则如果最远但是不在当前桶里那还是要遍历，否则最远且在桶里就可以更新当前已知最远点。 - // 目前先这样只管完全落在桶里的 + // //目前先这样只管完全落在桶里的 + // if (CONFIG.isAcc_convex() + // && chunkSuit4Tri.chunkMetadata.getStartTime() >= localCurStartTime + // && chunkSuit4Tri.chunkMetadata.getEndTime() < localCurEndTime + // && chunkSuit4Tri.chunkMetadata.getStatistics().getCount() >= 3 // 不考虑少于三个点 + // ) { + // BitSet bitSet = + // chunkSuit4Tri.chunkMetadata.getStatistics().getQuickHullBitSet(); + // List<QuickHullPoint> foundPoints = + // convexHullAcc( + // lt, + // lv, + // rt, + // rv, + // pageReader, + // bitSet, + // chunkSuit4Tri.chunkMetadata.getStatistics().getCount()); // + // 有可能不止两个点，当一边是平行线两端点 + // // System.out.println(foundPoints); + // for (QuickHullPoint point : foundPoints) { + // IOMonitor2.DCP_D_getAllSatisfiedPageData_traversedPointNum++; + // double distance = IOMonitor2.calculateDistance(lt, lv, point.t, point.v, + // rt, rv); + // if (distance > maxDistance) { + // // 是不是因为开启了acc_rect之后，导致这里要遍历的chunk块里没有点的距离可以达到maxDistance + // // 从而acc_convex不会生效？！ + // maxDistance = distance; + // select_t = point.t; + // select_v = point.v; + // } + // } + // continue; // note this + // } if (CONFIG.isAcc_convex() - && chunkSuit4Tri.chunkMetadata.getStartTime() >= localCurStartTime - && chunkSuit4Tri.chunkMetadata.getEndTime() < localCurEndTime && chunkSuit4Tri.chunkMetadata.getStatistics().getCount() >= 3 // 不考虑少于三个点 ) { BitSet bitSet = chunkSuit4Tri.chunkMetadata.getStatistics().getQuickHullBitSet(); @@ -401,18 +431,33 @@ public class LocalGroupByExecutorTri_ILTS implements GroupByExecutor { bitSet, chunkSuit4Tri.chunkMetadata.getStatistics().getCount()); // 有可能不止两个点，当一边是平行线两端点 // System.out.println(foundPoints); + double ch_maxDistance = -1; + long ch_select_t = -1; + double ch_select_v = -1; + // 找到foundPoints里的最远点 for (QuickHullPoint point : foundPoints) { IOMonitor2.DCP_D_getAllSatisfiedPageData_traversedPointNum++; double distance = IOMonitor2.calculateDistance(lt, lv, point.t, point.v, rt, rv); - if (distance > maxDistance) { + if (distance > ch_maxDistance) { // 是不是因为开启了acc_rect之后，导致这里要遍历的chunk块里没有点的距离可以达到maxDistance // 从而acc_convex不会生效？！ - maxDistance = distance; - select_t = point.t; - select_v = point.v; + ch_maxDistance = distance; + ch_select_t = point.t; + ch_select_v = point.v; } } - continue; // note this + // 和当前找到的最远距离比较 + if (ch_maxDistance <= maxDistance) { + continue; // 这个块里一定没有比当前找到的最远点更远的点，不管块凸包最远点在不在当前桶里都不用管了 + } + // 否则ch_maxDistance>maxDistance，还要判断落在当前桶内才行 + if (ch_select_t >= localCurStartTime && ch_select_t < localCurEndTime) { + maxDistance = ch_maxDistance; + select_t = ch_select_t; + select_v = ch_select_v; + continue; // note this + } + // 否则ch_maxDistance>maxDistance但是这个块的凸包最远点不在当前桶里，于是继续下面的遍历点操作 } int count = chunkSuit4Tri.chunkMetadata.getStatistics().getCount(); int j;
