indexes WIP
This commit is contained in:
82
index.h
82
index.h
@@ -1,79 +1,77 @@
|
||||
#pragma once
|
||||
|
||||
#include "exception.h"
|
||||
|
||||
#include <iostream>
|
||||
#include <utility>
|
||||
#include <vector>
|
||||
#include <map>
|
||||
|
||||
|
||||
namespace usql {
|
||||
|
||||
enum class IndexedDataType {
|
||||
integer,
|
||||
string
|
||||
};
|
||||
|
||||
using rowid_t = size_t; // int is now enough but size_t is correct
|
||||
|
||||
template <typename K>
|
||||
template<typename K>
|
||||
class Index {
|
||||
public:
|
||||
Index(std::string index_name, std::string col_name, IndexedDataType type) :
|
||||
m_index_name(std::move(index_name)), m_column_name(std::move(col_name)),
|
||||
m_data_type(type), m_uniq(false) {}
|
||||
|
||||
void insert(K key, int rowid) {
|
||||
// std::cout << "inserting key: " << key << " val: " << rowid << std::endl;
|
||||
|
||||
// TODO handle uniqueness
|
||||
void insert(K key, rowid_t rowid) {
|
||||
auto search = m_index.find(key);
|
||||
if (search != m_index.end()) {
|
||||
if (m_uniq)
|
||||
throw Exception("Inserting duplicate value into unique index");
|
||||
|
||||
search->second.push_back(rowid);
|
||||
} else {
|
||||
std::vector<int> rowids{rowid};
|
||||
std::vector<rowid_t> rowids{rowid};
|
||||
if (!m_uniq)
|
||||
rowids.reserve(8);
|
||||
m_index[key] = rowids;
|
||||
}
|
||||
}
|
||||
}
|
||||
|
||||
void update(K old_key, K new_key, int rowid) {
|
||||
// std::cout << "updating key: " << old_key << " to: " << new_key << " val: " << rowid << std::endl;
|
||||
|
||||
// TODO handle uniqueness
|
||||
}
|
||||
|
||||
void remove(K key, int rowid) {
|
||||
// std::cout << "removing key: " << key << " val: " << rowid << std::endl;
|
||||
|
||||
void remove(K key, rowid_t rowid) {
|
||||
auto search = m_index.find(key);
|
||||
if (search != m_index.end()) {
|
||||
search->second.erase(find(search->second.begin(), search->second.end(), rowid));
|
||||
}
|
||||
if (search->second.empty())
|
||||
m_index.erase(search);
|
||||
}
|
||||
}
|
||||
|
||||
std::vector<int> search(K key) {
|
||||
// std::cout << "returning rowids for key: " << key << std::endl;
|
||||
|
||||
std::vector<rowid_t> search(K key) {
|
||||
auto search = m_index.find(key);
|
||||
if (search != m_index.end()) {
|
||||
return search->second;
|
||||
} else {
|
||||
return std::vector<int>{};
|
||||
}
|
||||
return std::vector<rowid_t>{};
|
||||
}
|
||||
}
|
||||
|
||||
void truncate() {
|
||||
// std::cout << "truncating" << std::endl;
|
||||
m_index.clear();
|
||||
}
|
||||
|
||||
void dump() {
|
||||
std::for_each(m_index.begin(), m_index.end(),
|
||||
[](std::pair<K, std::vector<int>> element){
|
||||
K key = element.first;
|
||||
std::vector<int> rowids = element.second;
|
||||
std::cout << "key: " << key << ", rowids count:" << rowids.size() << std::endl;
|
||||
});
|
||||
}
|
||||
// void dump() {
|
||||
// std::for_each(m_index.begin(), m_index.end(),
|
||||
// [](std::pair<K, std::vector<rowid_t>> element){
|
||||
// K key = element.first;
|
||||
// std::vector<rowid_t> rowids = element.second;
|
||||
// std::cout << "key: " << key << ", rowids count:" << rowids.size() << std::endl;
|
||||
// });
|
||||
// }
|
||||
|
||||
[[nodiscard]] const std::string &get_column_name() const {
|
||||
return m_column_name;
|
||||
return m_column_name;
|
||||
}
|
||||
|
||||
[[nodiscard]] const std::string &get_index_name() const {
|
||||
@@ -83,14 +81,14 @@ public:
|
||||
[[nodiscard]] IndexedDataType get_data_type() const {
|
||||
return m_data_type;
|
||||
}
|
||||
|
||||
private:
|
||||
bool m_uniq;
|
||||
std::string m_index_name;
|
||||
std::string m_column_name;
|
||||
IndexedDataType m_data_type;
|
||||
|
||||
// DEBUG for debug it is public
|
||||
public:
|
||||
std::map<K, std::vector<int> > m_index;
|
||||
};
|
||||
private:
|
||||
bool m_uniq;
|
||||
std::string m_index_name;
|
||||
std::string m_column_name;
|
||||
IndexedDataType m_data_type;
|
||||
|
||||
std::map<K, std::vector<rowid_t> > m_index;
|
||||
};
|
||||
|
||||
} // namespace
|
||||
Reference in New Issue
Block a user