ClickHouse/src/Interpreters/TranslateQualifiedNamesVisitor.cpp

372 lines
13 KiB
C++
Raw Normal View History

#include <Poco/String.h>
#include <Interpreters/TranslateQualifiedNamesVisitor.h>
#include <Interpreters/IdentifierSemantic.h>
#include <Common/typeid_cast.h>
2020-04-12 03:20:15 +00:00
#include <Common/StringUtils/StringUtils.h>
2018-11-02 18:53:23 +00:00
#include <Core/Names.h>
2018-09-28 15:01:13 +00:00
#include <Parsers/ASTIdentifier.h>
#include <Parsers/ASTAsterisk.h>
#include <Parsers/ASTQualifiedAsterisk.h>
#include <Parsers/ASTSelectQuery.h>
#include <Parsers/ASTSelectWithUnionQuery.h>
#include <Parsers/ASTTablesInSelectQuery.h>
#include <Parsers/ASTExpressionList.h>
#include <Parsers/ASTLiteral.h>
#include <Parsers/ASTFunction.h>
2019-07-21 17:03:58 +00:00
#include <Parsers/ASTColumnsMatcher.h>
2020-08-29 05:33:46 +00:00
#include <Parsers/ASTColumnsTransformers.h>
2018-11-02 18:53:23 +00:00
namespace DB
{
namespace ErrorCodes
{
extern const int UNKNOWN_IDENTIFIER;
2020-06-03 10:50:15 +00:00
extern const int UNSUPPORTED_JOIN_KEYS;
extern const int LOGICAL_ERROR;
}
2019-11-13 16:49:29 +00:00
bool TranslateQualifiedNamesMatcher::Data::unknownColumn(size_t table_pos, const ASTIdentifier & identifier) const
{
const auto & table = tables[table_pos].table;
2019-11-13 16:49:29 +00:00
auto nested1 = IdentifierSemantic::extractNestedName(identifier, table.table);
auto nested2 = IdentifierSemantic::extractNestedName(identifier, table.alias);
2020-03-18 03:27:32 +00:00
const String & short_name = identifier.shortName();
const auto & columns = tables[table_pos].columns;
for (const auto & column : columns)
2019-11-13 16:49:29 +00:00
{
const String & known_name = column.name;
2019-11-13 16:49:29 +00:00
if (short_name == known_name)
return false;
if (nested1 && *nested1 == known_name)
return false;
if (nested2 && *nested2 == known_name)
return false;
}
const auto & hidden_columns = tables[table_pos].hidden_columns;
for (const auto & column : hidden_columns)
{
const String & known_name = column.name;
if (short_name == known_name)
return false;
if (nested1 && *nested1 == known_name)
return false;
if (nested2 && *nested2 == known_name)
return false;
}
return !columns.empty();
2019-11-13 16:49:29 +00:00
}
bool TranslateQualifiedNamesMatcher::needChildVisit(ASTPtr & node, const ASTPtr & child)
{
/// Do not go to FROM, JOIN, subqueries.
2019-03-11 13:22:51 +00:00
if (child->as<ASTTableExpression>() || child->as<ASTSelectWithUnionQuery>())
return false;
/// Processed nodes. Do not go into children.
2019-03-11 13:22:51 +00:00
if (node->as<ASTQualifiedAsterisk>() || node->as<ASTTableJoin>())
return false;
/// ASTSelectQuery + others
return true;
}
2019-02-22 13:33:56 +00:00
void TranslateQualifiedNamesMatcher::visit(ASTPtr & ast, Data & data)
{
2019-03-11 13:22:51 +00:00
if (auto * t = ast->as<ASTIdentifier>())
2019-02-22 13:33:56 +00:00
visit(*t, ast, data);
2019-03-11 13:22:51 +00:00
if (auto * t = ast->as<ASTTableJoin>())
2019-02-22 13:33:56 +00:00
visit(*t, ast, data);
2019-03-11 13:22:51 +00:00
if (auto * t = ast->as<ASTSelectQuery>())
2019-02-22 13:33:56 +00:00
visit(*t, ast, data);
2019-03-11 13:22:51 +00:00
if (auto * node = ast->as<ASTExpressionList>())
visit(*node, ast, data);
2019-03-11 13:22:51 +00:00
if (auto * node = ast->as<ASTFunction>())
visit(*node, ast, data);
}
2019-02-22 13:33:56 +00:00
void TranslateQualifiedNamesMatcher::visit(ASTIdentifier & identifier, ASTPtr &, Data & data)
{
if (IdentifierSemantic::getColumnName(identifier))
{
String short_name = identifier.shortName();
bool allow_ambiguous = data.join_using_columns.count(short_name);
if (auto best_pos = IdentifierSemantic::chooseTable(identifier, data.tables, allow_ambiguous))
{
size_t table_pos = *best_pos;
2019-11-13 16:49:29 +00:00
if (data.unknownColumn(table_pos, identifier))
{
String table_name = data.tables[table_pos].table.getQualifiedNamePrefix(false);
throw Exception("There's no column '" + identifier.name() + "' in table '" + table_name + "'",
ErrorCodes::UNKNOWN_IDENTIFIER);
}
IdentifierSemantic::setMembership(identifier, table_pos);
/// In case if column from the joined table are in source columns, change it's name to qualified.
2020-04-12 03:20:15 +00:00
/// Also always leave unusual identifiers qualified.
2020-04-22 06:01:33 +00:00
const auto & table = data.tables[table_pos].table;
2020-04-12 03:20:15 +00:00
if (table_pos && (data.hasColumn(short_name) || !isValidIdentifierBegin(short_name.at(0))))
IdentifierSemantic::setColumnLongName(identifier, table);
else
IdentifierSemantic::setColumnShortName(identifier, table);
}
}
}
/// As special case, treat count(*) as count(), not as count(list of all columns).
void TranslateQualifiedNamesMatcher::visit(ASTFunction & node, const ASTPtr &, Data &)
{
ASTPtr & func_arguments = node.arguments;
if (!func_arguments) return;
String func_name_lowercase = Poco::toLower(node.name);
if (func_name_lowercase == "count" &&
func_arguments->children.size() == 1 &&
2019-03-11 13:22:51 +00:00
func_arguments->children[0]->as<ASTAsterisk>())
func_arguments->children.clear();
}
2020-04-12 03:20:15 +00:00
void TranslateQualifiedNamesMatcher::visit(const ASTQualifiedAsterisk &, const ASTPtr & ast, Data & data)
{
2020-08-29 05:33:46 +00:00
if (ast->children.empty())
throw Exception("Logical error: qualified asterisk must have children", ErrorCodes::LOGICAL_ERROR);
2019-01-16 12:08:43 +00:00
auto & ident = ast->children[0];
2019-01-16 12:08:43 +00:00
/// @note it could contain table alias as table name.
DatabaseAndTableWithAlias db_and_table(ident);
2019-01-16 12:08:43 +00:00
for (const auto & known_table : data.tables)
if (db_and_table.satisfies(known_table.table, true))
2019-02-22 13:33:56 +00:00
return;
throw Exception("Unknown qualified identifier: " + ident->getAliasOrColumnName(), ErrorCodes::UNKNOWN_IDENTIFIER);
}
2019-02-22 13:33:56 +00:00
void TranslateQualifiedNamesMatcher::visit(ASTTableJoin & join, const ASTPtr & , Data & data)
{
if (join.using_expression_list)
Visitor(data).visit(join.using_expression_list);
else if (join.on_expression)
Visitor(data).visit(join.on_expression);
}
2019-02-22 13:33:56 +00:00
void TranslateQualifiedNamesMatcher::visit(ASTSelectQuery & select, const ASTPtr & , Data & data)
{
2020-04-22 06:01:33 +00:00
if (const auto * join = select.join())
extractJoinUsingColumns(join->table_join, data);
2019-04-09 14:59:06 +00:00
2019-01-22 19:56:53 +00:00
/// If the WHERE clause or HAVING consists of a single qualified column, the reference must be translated not only in children,
/// but also in where_expression and having_expression.
if (select.prewhere())
2019-04-09 14:59:06 +00:00
Visitor(data).visit(select.refPrewhere());
if (select.where())
2019-04-09 14:59:06 +00:00
Visitor(data).visit(select.refWhere());
if (select.having())
2019-04-09 14:59:06 +00:00
Visitor(data).visit(select.refHaving());
}
static void addIdentifier(ASTs & nodes, const DatabaseAndTableWithAlias & table, const String & column_name)
{
std::vector<String> parts = {column_name};
String table_name = table.getQualifiedNamePrefix(false);
if (!table_name.empty()) parts.insert(parts.begin(), table_name);
nodes.emplace_back(std::make_shared<ASTIdentifier>(std::move(parts)));
}
/// Replace *, alias.*, database.table.* with a list of columns.
void TranslateQualifiedNamesMatcher::visit(ASTExpressionList & node, const ASTPtr &, Data & data)
{
const auto & tables_with_columns = data.tables;
ASTs old_children;
if (data.processAsterisks())
{
bool has_asterisk = false;
for (const auto & child : node.children)
{
2019-07-21 17:03:58 +00:00
if (child->as<ASTAsterisk>() || child->as<ASTColumnsMatcher>())
{
if (tables_with_columns.empty())
throw Exception("An asterisk cannot be replaced with empty columns.", ErrorCodes::LOGICAL_ERROR);
has_asterisk = true;
}
2019-03-11 13:22:51 +00:00
else if (const auto * qa = child->as<ASTQualifiedAsterisk>())
{
visit(*qa, child, data); /// check if it's OK before rewrite
has_asterisk = true;
}
}
if (has_asterisk)
{
old_children.swap(node.children);
node.children.reserve(old_children.size());
}
}
for (const auto & child : old_children)
{
2020-09-28 09:07:57 +00:00
ASTs columns;
2019-03-11 13:22:51 +00:00
if (const auto * asterisk = child->as<ASTAsterisk>())
{
bool first_table = true;
for (const auto & table : tables_with_columns)
{
for (const auto * cols : {&table.columns, &table.alias_columns, &table.materialized_columns})
{
for (const auto & column : *cols)
{
if (first_table || !data.join_using_columns.count(column.name))
{
addIdentifier(columns, table.table, column.name);
}
}
}
first_table = false;
}
2020-08-29 05:33:46 +00:00
for (const auto & transformer : asterisk->children)
{
2020-09-28 09:07:57 +00:00
IASTColumnsTransformer::transform(transformer, columns);
2020-08-29 05:33:46 +00:00
}
}
2019-07-21 17:03:58 +00:00
else if (const auto * asterisk_pattern = child->as<ASTColumnsMatcher>())
2019-07-12 11:17:38 +00:00
{
if (asterisk_pattern->column_list)
2019-07-12 11:17:38 +00:00
{
for (const auto & ident : asterisk_pattern->column_list->children)
2020-09-28 09:07:57 +00:00
columns.emplace_back(ident->clone());
}
else
{
bool first_table = true;
for (const auto & table : tables_with_columns)
2019-07-12 11:17:38 +00:00
{
for (const auto & column : table.columns)
2019-07-12 11:17:38 +00:00
{
if (asterisk_pattern->isColumnMatching(column.name) && (first_table || !data.join_using_columns.count(column.name)))
{
2020-10-03 13:54:23 +00:00
addIdentifier(columns, table.table, column.name);
}
2019-07-12 11:17:38 +00:00
}
first_table = false;
}
2019-07-12 11:17:38 +00:00
}
2020-08-29 05:33:46 +00:00
// ColumnsMatcher's transformers start to appear at child 1
for (auto it = asterisk_pattern->children.begin() + 1; it != asterisk_pattern->children.end(); ++it)
{
2020-09-28 09:07:57 +00:00
IASTColumnsTransformer::transform(*it, columns);
2020-08-29 05:33:46 +00:00
}
2019-07-12 11:17:38 +00:00
}
2019-03-11 13:22:51 +00:00
else if (const auto * qualified_asterisk = child->as<ASTQualifiedAsterisk>())
{
DatabaseAndTableWithAlias ident_db_and_name(qualified_asterisk->children[0]);
for (const auto & table : tables_with_columns)
{
if (ident_db_and_name.satisfies(table.table, true))
{
for (const auto & column : table.columns)
{
2020-10-03 13:54:23 +00:00
addIdentifier(columns, table.table, column.name);
}
break;
}
}
2020-08-29 05:33:46 +00:00
// QualifiedAsterisk's transformers start to appear at child 1
for (auto it = qualified_asterisk->children.begin() + 1; it != qualified_asterisk->children.end(); ++it)
{
2020-09-28 09:07:57 +00:00
IASTColumnsTransformer::transform(*it, columns);
2020-08-29 05:33:46 +00:00
}
}
else
2020-09-28 09:07:57 +00:00
columns.emplace_back(child);
node.children.insert(
node.children.end(),
std::make_move_iterator(columns.begin()),
std::make_move_iterator(columns.end()));
}
}
/// 'select * from a join b using id' should result one 'id' column
void TranslateQualifiedNamesMatcher::extractJoinUsingColumns(const ASTPtr ast, Data & data)
{
const auto & table_join = ast->as<ASTTableJoin &>();
if (table_join.using_expression_list)
{
const auto & keys = table_join.using_expression_list->as<ASTExpressionList &>();
for (const auto & key : keys.children)
2019-08-08 20:02:30 +00:00
if (auto opt_column = tryGetIdentifierName(key))
data.join_using_columns.insert(*opt_column);
2019-03-11 13:22:51 +00:00
else if (key->as<ASTLiteral>())
data.join_using_columns.insert(key->getColumnName());
else
{
String alias = key->tryGetAlias();
if (alias.empty())
2020-06-03 10:50:15 +00:00
throw Exception("Wrong key in USING. Expected identifier or alias, got: " + key->getID(),
ErrorCodes::UNSUPPORTED_JOIN_KEYS);
data.join_using_columns.insert(alias);
}
}
}
void RestoreQualifiedNamesMatcher::Data::changeTable(ASTIdentifier & identifier) const
{
auto match = IdentifierSemantic::canReferColumnToTable(identifier, distributed_table);
switch (match)
{
case IdentifierSemantic::ColumnMatch::AliasedTableName:
case IdentifierSemantic::ColumnMatch::TableName:
case IdentifierSemantic::ColumnMatch::DBAndTable:
IdentifierSemantic::setColumnLongName(identifier, remote_table);
break;
default:
break;
}
}
bool RestoreQualifiedNamesMatcher::needChildVisit(ASTPtr &, const ASTPtr & child)
{
/// Do not go into subqueries
if (child->as<ASTSelectWithUnionQuery>())
return false; // NOLINT
return true;
}
void RestoreQualifiedNamesMatcher::visit(ASTPtr & ast, Data & data)
{
if (auto * t = ast->as<ASTIdentifier>())
visit(*t, ast, data);
}
void RestoreQualifiedNamesMatcher::visit(ASTIdentifier & identifier, ASTPtr &, Data & data)
{
if (IdentifierSemantic::getColumnName(identifier))
{
2019-10-21 11:22:22 +00:00
if (IdentifierSemantic::getMembership(identifier))
{
identifier.restoreTable(); // TODO(ilezhankin): should restore qualified name here - why exactly here?
data.changeTable(identifier);
}
}
}
}