| /* |
| * Copyright (C) 2019 The Android Open Source Project |
| * |
| * Licensed under the Apache License, Version 2.0 (the "License"); |
| * you may not use this file except in compliance with the License. |
| * You may obtain a copy of the License at |
| * |
| * http://www.apache.org/licenses/LICENSE-2.0 |
| * |
| * Unless required by applicable law or agreed to in writing, software |
| * distributed under the License is distributed on an "AS IS" BASIS, |
| * WITHOUT WARRANTIES OR CONDITIONS OF ANY KIND, either express or implied. |
| * See the License for the specific language governing permissions and |
| * limitations under the License. |
| */ |
| |
| #include "src/trace_processor/db/table.h" |
| |
| namespace perfetto { |
| namespace trace_processor { |
| |
| Table::Table() = default; |
| Table::~Table() = default; |
| |
| Table::Table(StringPool* pool, const Table* parent) : string_pool_(pool) { |
| if (!parent) |
| return; |
| |
| // If this table has a parent, then copy over all the columns pointing to |
| // empty RowMaps. |
| for (uint32_t i = 0; i < parent->row_maps_.size(); ++i) |
| row_maps_.emplace_back(); |
| for (const Column& col : parent->columns_) |
| columns_.emplace_back(col, this, columns_.size(), col.row_map_idx_); |
| } |
| |
| Table& Table::operator=(Table&& other) noexcept { |
| row_count_ = other.row_count_; |
| string_pool_ = other.string_pool_; |
| |
| row_maps_ = std::move(other.row_maps_); |
| columns_ = std::move(other.columns_); |
| for (Column& col : columns_) { |
| col.table_ = this; |
| } |
| return *this; |
| } |
| |
| Table Table::Copy() const { |
| Table table = CopyExceptRowMaps(); |
| for (const RowMap& rm : row_maps_) { |
| table.row_maps_.emplace_back(rm.Copy()); |
| } |
| return table; |
| } |
| |
| Table Table::CopyExceptRowMaps() const { |
| Table table(string_pool_, nullptr); |
| table.row_count_ = row_count_; |
| for (const Column& col : columns_) { |
| table.columns_.emplace_back(col, &table, col.index_in_table(), |
| col.row_map_idx_); |
| } |
| return table; |
| } |
| |
| Table Table::Sort(const std::vector<Order>& od) const { |
| if (od.empty()) |
| return Copy(); |
| |
| // Return a copy if there is a single constraint to sort the table |
| // by a column which is already sorted. |
| const auto& first_col = GetColumn(od.front().col_idx); |
| if (od.size() == 1 && first_col.IsSorted() && !od.front().desc) |
| return Copy(); |
| |
| // Build an index vector with all the indices for the first |size_| rows. |
| std::vector<uint32_t> idx(row_count_); |
| |
| if (od.size() == 1 && first_col.IsSorted()) { |
| // We special case a single constraint in descending order as this |
| // happens any time the |max| function is used in SQLite. We can be |
| // more efficient as this column is already sorted so we simply need |
| // to reverse the order of this column. |
| PERFETTO_DCHECK(od.front().desc); |
| std::iota(idx.rbegin(), idx.rend(), 0); |
| } else { |
| // As our data is columnar, it's always more efficient to sort one column |
| // at a time rather than try and sort lexiographically all at once. |
| // To preserve correctness, we need to stably sort the index vector once |
| // for each order by in *reverse* order. Reverse order is important as it |
| // preserves the lexiographical property. |
| // |
| // For example, suppose we have the following: |
| // Table { |
| // Column x; |
| // Column y |
| // Column z; |
| // } |
| // |
| // Then, to sort "y asc, x desc", we could do one of two things: |
| // 1) sort the index vector all at once and on each index, we compare |
| // y then z. This is slow as the data is columnar and we need to |
| // repeatedly branch inside each column. |
| // 2) we can stably sort first on x desc and then sort on y asc. This will |
| // first put all the x in the correct order such that when we sort on |
| // y asc, we will have the correct order of x where y is the same (since |
| // the sort is stable). |
| // |
| // TODO(lalitm): it is possible that we could sort the last constraint (i.e. |
| // the first constraint in the below loop) in a non-stable way. However, |
| // this is more subtle than it appears as we would then need special |
| // handling where there are order bys on a column which is already sorted |
| // (e.g. ts, id). Investigate whether the performance gains from this are |
| // worthwhile. This also needs changes to the constraint modification logic |
| // in DbSqliteTable which currently eliminates constraints on sorted |
| // columns. |
| std::iota(idx.begin(), idx.end(), 0); |
| for (auto it = od.rbegin(); it != od.rend(); ++it) { |
| columns_[it->col_idx].StableSort(it->desc, &idx); |
| } |
| } |
| |
| // Return a copy of this table with the RowMaps using the computed ordered |
| // RowMap. |
| Table table = CopyExceptRowMaps(); |
| RowMap rm(std::move(idx)); |
| for (const RowMap& map : row_maps_) { |
| table.row_maps_.emplace_back(map.SelectRows(rm)); |
| PERFETTO_DCHECK(table.row_maps_.back().size() == table.row_count()); |
| } |
| |
| // Remove the sorted flag from all the columns. |
| for (auto& col : table.columns_) { |
| col.flags_ &= ~Column::Flag::kSorted; |
| } |
| |
| // For the first order by, make the column flag itself as sorted but |
| // only if the sort was in ascending order. |
| if (!od.front().desc) { |
| table.columns_[od.front().col_idx].flags_ |= Column::Flag::kSorted; |
| } |
| |
| return table; |
| } |
| |
| Table Table::LookupJoin(JoinKey left, const Table& other, JoinKey right) { |
| // The join table will have the same size and RowMaps as the left (this) |
| // table because the left column is indexing the right table. |
| Table table(string_pool_, nullptr); |
| table.row_count_ = row_count_; |
| for (const RowMap& rm : row_maps_) { |
| table.row_maps_.emplace_back(rm.Copy()); |
| } |
| |
| for (const Column& col : columns_) { |
| // We skip id columns as they are misleading on join tables. |
| if (col.IsId()) |
| continue; |
| table.columns_.emplace_back(col, &table, table.columns_.size(), |
| col.row_map_idx_); |
| } |
| |
| const Column& left_col = columns_[left.col_idx]; |
| const Column& right_col = other.columns_[right.col_idx]; |
| |
| // For each index in the left column, retrieve the index of the row inside |
| // the RowMap of the right column. By getting the index of the row rather |
| // than the row number itself, we can call |Apply| on the other RowMaps |
| // in the right table. |
| std::vector<uint32_t> indices(row_count_); |
| for (uint32_t i = 0; i < row_count_; ++i) { |
| SqlValue val = left_col.Get(i); |
| PERFETTO_CHECK(val.type != SqlValue::Type::kNull); |
| indices[i] = right_col.IndexOf(val).value(); |
| } |
| |
| // Apply the computed RowMap to each of the right RowMaps, adding it to the |
| // join table as we go. |
| RowMap rm(std::move(indices)); |
| for (const RowMap& ot : other.row_maps_) { |
| table.row_maps_.emplace_back(ot.SelectRows(rm)); |
| } |
| |
| uint32_t left_row_maps_size = static_cast<uint32_t>(row_maps_.size()); |
| for (const Column& col : other.columns_) { |
| // We skip id columns as they are misleading on join tables. |
| if (col.IsId()) |
| continue; |
| |
| // Ensure that we offset the RowMap index by the number of RowMaps in the |
| // left table. |
| table.columns_.emplace_back(col, &table, table.columns_.size(), |
| col.row_map_idx_ + left_row_maps_size); |
| } |
| return table; |
| } |
| |
| } // namespace trace_processor |
| } // namespace perfetto |