diff --git a/src/Functions/FunctionStringToString.h b/src/Functions/FunctionStringToString.h index 62d7b09f79e..e0e64e47b49 100644 --- a/src/Functions/FunctionStringToString.h +++ b/src/Functions/FunctionStringToString.h @@ -59,19 +59,19 @@ public: bool useDefaultImplementationForConstants() const override { return true; } - ColumnPtr executeImpl(const ColumnsWithTypeAndName & arguments, const DataTypePtr &, size_t /*input_rows_count*/) const override + ColumnPtr executeImpl(const ColumnsWithTypeAndName & arguments, const DataTypePtr &, size_t input_rows_count) const override { const ColumnPtr column = arguments[0].column; if (const ColumnString * col = checkAndGetColumn(column.get())) { auto col_res = ColumnString::create(); - Impl::vector(col->getChars(), col->getOffsets(), col_res->getChars(), col_res->getOffsets()); + Impl::vector(col->getChars(), col->getOffsets(), col_res->getChars(), col_res->getOffsets(), input_rows_count); return col_res; } else if (const ColumnFixedString * col_fixed = checkAndGetColumn(column.get())) { auto col_res = ColumnFixedString::create(col_fixed->getN()); - Impl::vectorFixed(col_fixed->getChars(), col_fixed->getN(), col_res->getChars()); + Impl::vectorFixed(col_fixed->getChars(), col_fixed->getN(), col_res->getChars(), input_rows_count); return col_res; } else diff --git a/src/Functions/LowerUpperImpl.h b/src/Functions/LowerUpperImpl.h index 72b3ce1ca34..d463ef96e16 100644 --- a/src/Functions/LowerUpperImpl.h +++ b/src/Functions/LowerUpperImpl.h @@ -8,17 +8,19 @@ namespace DB template struct LowerUpperImpl { - static void vector(const ColumnString::Chars & data, + static void vector( + const ColumnString::Chars & data, const ColumnString::Offsets & offsets, ColumnString::Chars & res_data, - ColumnString::Offsets & res_offsets) + ColumnString::Offsets & res_offsets, + size_t /*input_rows_count*/) { res_data.resize_exact(data.size()); res_offsets.assign(offsets); array(data.data(), data.data() + data.size(), res_data.data()); } - static void vectorFixed(const ColumnString::Chars & data, size_t /*n*/, ColumnString::Chars & res_data) + static void vectorFixed(const ColumnString::Chars & data, size_t /*n*/, ColumnString::Chars & res_data, size_t /*input_rows_count*/) { res_data.resize_exact(data.size()); array(data.data(), data.data() + data.size(), res_data.data()); diff --git a/src/Functions/LowerUpperUTF8Impl.h b/src/Functions/LowerUpperUTF8Impl.h index eebba7b9d5f..eedabca5b22 100644 --- a/src/Functions/LowerUpperUTF8Impl.h +++ b/src/Functions/LowerUpperUTF8Impl.h @@ -90,7 +90,8 @@ struct LowerUpperUTF8Impl const ColumnString::Chars & data, const ColumnString::Offsets & offsets, ColumnString::Chars & res_data, - ColumnString::Offsets & res_offsets) + ColumnString::Offsets & res_offsets, + size_t input_rows_count) { if (data.empty()) return; @@ -98,7 +99,7 @@ struct LowerUpperUTF8Impl bool all_ascii = isAllASCII(data.data(), data.size()); if (all_ascii) { - LowerUpperImpl::vector(data, offsets, res_data, res_offsets); + LowerUpperImpl::vector(data, offsets, res_data, res_offsets, input_rows_count); return; } @@ -107,7 +108,7 @@ struct LowerUpperUTF8Impl array(data.data(), data.data() + data.size(), offsets, res_data.data()); } - static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &) + static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &, size_t) { throw Exception(ErrorCodes::BAD_ARGUMENTS, "Functions lowerUTF8 and upperUTF8 cannot work with FixedString argument"); } diff --git a/src/Functions/StringHelpers.h b/src/Functions/StringHelpers.h index 8f3a87d5d0e..24d8fe86c62 100644 --- a/src/Functions/StringHelpers.h +++ b/src/Functions/StringHelpers.h @@ -62,12 +62,13 @@ using Pos = const char *; template struct ExtractSubstringImpl { - static void vector(const ColumnString::Chars & data, const ColumnString::Offsets & offsets, - ColumnString::Chars & res_data, ColumnString::Offsets & res_offsets) + static void vector( + const ColumnString::Chars & data, const ColumnString::Offsets & offsets, + ColumnString::Chars & res_data, ColumnString::Offsets & res_offsets, + size_t input_rows_count) { - size_t size = offsets.size(); - res_offsets.resize(size); - res_data.reserve(size * Extractor::getReserveLengthForElement()); + res_offsets.resize(input_rows_count); + res_data.reserve(input_rows_count * Extractor::getReserveLengthForElement()); size_t prev_offset = 0; size_t res_offset = 0; @@ -76,7 +77,7 @@ struct ExtractSubstringImpl Pos start; size_t length; - for (size_t i = 0; i < size; ++i) + for (size_t i = 0; i < input_rows_count; ++i) { Extractor::execute(reinterpret_cast(&data[prev_offset]), offsets[i] - prev_offset - 1, start, length); @@ -99,7 +100,7 @@ struct ExtractSubstringImpl res_data.assign(start, length); } - static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &) + static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &, size_t) { throw Exception(ErrorCodes::ILLEGAL_COLUMN, "Column of type FixedString is not supported by this function"); } @@ -111,12 +112,13 @@ struct ExtractSubstringImpl template struct CutSubstringImpl { - static void vector(const ColumnString::Chars & data, const ColumnString::Offsets & offsets, - ColumnString::Chars & res_data, ColumnString::Offsets & res_offsets) + static void vector( + const ColumnString::Chars & data, const ColumnString::Offsets & offsets, + ColumnString::Chars & res_data, ColumnString::Offsets & res_offsets, + size_t input_rows_count) { res_data.reserve(data.size()); - size_t size = offsets.size(); - res_offsets.resize(size); + res_offsets.resize(input_rows_count); size_t prev_offset = 0; size_t res_offset = 0; @@ -125,7 +127,7 @@ struct CutSubstringImpl Pos start; size_t length; - for (size_t i = 0; i < size; ++i) + for (size_t i = 0; i < input_rows_count; ++i) { const char * current = reinterpret_cast(&data[prev_offset]); Extractor::execute(current, offsets[i] - prev_offset - 1, start, length); @@ -154,7 +156,7 @@ struct CutSubstringImpl res_data.append(start + length, data.data() + data.size()); } - static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &) + static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &, size_t) { throw Exception(ErrorCodes::ILLEGAL_COLUMN, "Column of type FixedString is not supported by this function"); } diff --git a/src/Functions/URL/ExtractFirstSignificantSubdomain.h b/src/Functions/URL/ExtractFirstSignificantSubdomain.h index 0d1b1cac8ef..4316fd2fc3a 100644 --- a/src/Functions/URL/ExtractFirstSignificantSubdomain.h +++ b/src/Functions/URL/ExtractFirstSignificantSubdomain.h @@ -1,8 +1,8 @@ #pragma once #include -#include "domain.h" -#include "tldLookup.h" +#include +#include #include /// TLDType namespace DB diff --git a/src/Functions/URL/cutFragment.cpp b/src/Functions/URL/cutFragment.cpp index 3b99edf61a3..b32c4410190 100644 --- a/src/Functions/URL/cutFragment.cpp +++ b/src/Functions/URL/cutFragment.cpp @@ -1,6 +1,6 @@ #include -#include "fragment.h" #include +#include namespace DB { diff --git a/src/Functions/URL/cutQueryString.cpp b/src/Functions/URL/cutQueryString.cpp index 2886adc2e36..d2fa4a004f2 100644 --- a/src/Functions/URL/cutQueryString.cpp +++ b/src/Functions/URL/cutQueryString.cpp @@ -1,6 +1,6 @@ #include -#include "queryString.h" #include +#include namespace DB { diff --git a/src/Functions/URL/cutQueryStringAndFragment.cpp b/src/Functions/URL/cutQueryStringAndFragment.cpp index 4116b352542..ff427beb277 100644 --- a/src/Functions/URL/cutQueryStringAndFragment.cpp +++ b/src/Functions/URL/cutQueryStringAndFragment.cpp @@ -1,6 +1,6 @@ #include -#include "queryStringAndFragment.h" #include +#include namespace DB { diff --git a/src/Functions/URL/cutToFirstSignificantSubdomain.cpp b/src/Functions/URL/cutToFirstSignificantSubdomain.cpp index 6e64b0b6ab8..454a241c54f 100644 --- a/src/Functions/URL/cutToFirstSignificantSubdomain.cpp +++ b/src/Functions/URL/cutToFirstSignificantSubdomain.cpp @@ -1,6 +1,6 @@ #include #include -#include "ExtractFirstSignificantSubdomain.h" +#include namespace DB diff --git a/src/Functions/URL/cutToFirstSignificantSubdomainCustom.cpp b/src/Functions/URL/cutToFirstSignificantSubdomainCustom.cpp index 77f40e465a6..7612b6ea4af 100644 --- a/src/Functions/URL/cutToFirstSignificantSubdomainCustom.cpp +++ b/src/Functions/URL/cutToFirstSignificantSubdomainCustom.cpp @@ -1,6 +1,6 @@ #include -#include "ExtractFirstSignificantSubdomain.h" -#include "FirstSignificantSubdomainCustomImpl.h" +#include +#include namespace DB { diff --git a/src/Functions/URL/cutWWW.cpp b/src/Functions/URL/cutWWW.cpp index 992d5128440..ab3fae6b094 100644 --- a/src/Functions/URL/cutWWW.cpp +++ b/src/Functions/URL/cutWWW.cpp @@ -1,6 +1,6 @@ #include #include -#include "protocol.h" +#include #include diff --git a/src/Functions/URL/decodeURLComponent.cpp b/src/Functions/URL/decodeURLComponent.cpp index 05e3fbea3fd..bf4aaa6d5e3 100644 --- a/src/Functions/URL/decodeURLComponent.cpp +++ b/src/Functions/URL/decodeURLComponent.cpp @@ -1,7 +1,7 @@ -#include #include #include #include +#include namespace DB @@ -121,8 +121,10 @@ enum URLCodeStrategy template struct CodeURLComponentImpl { - static void vector(const ColumnString::Chars & data, const ColumnString::Offsets & offsets, - ColumnString::Chars & res_data, ColumnString::Offsets & res_offsets) + static void vector( + const ColumnString::Chars & data, const ColumnString::Offsets & offsets, + ColumnString::Chars & res_data, ColumnString::Offsets & res_offsets, + size_t input_rows_count) { if (code_strategy == encode) { @@ -134,13 +136,12 @@ struct CodeURLComponentImpl res_data.resize(data.size()); } - size_t size = offsets.size(); - res_offsets.resize(size); + res_offsets.resize(input_rows_count); size_t prev_offset = 0; size_t res_offset = 0; - for (size_t i = 0; i < size; ++i) + for (size_t i = 0; i < input_rows_count; ++i) { const char * src_data = reinterpret_cast(&data[prev_offset]); size_t src_size = offsets[i] - prev_offset; @@ -165,7 +166,7 @@ struct CodeURLComponentImpl res_data.resize(res_offset); } - [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &) + [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &, size_t) { throw Exception(ErrorCodes::ILLEGAL_COLUMN, "Column of type FixedString is not supported by URL functions"); } diff --git a/src/Functions/URL/domain.cpp b/src/Functions/URL/domain.cpp index 443a3323075..68724b57a70 100644 --- a/src/Functions/URL/domain.cpp +++ b/src/Functions/URL/domain.cpp @@ -1,5 +1,4 @@ -#include "domain.h" - +#include #include #include diff --git a/src/Functions/URL/domain.h b/src/Functions/URL/domain.h index 936fb9d5f00..328df76b570 100644 --- a/src/Functions/URL/domain.h +++ b/src/Functions/URL/domain.h @@ -1,9 +1,10 @@ #pragma once -#include "protocol.h" -#include -#include #include +#include +#include + +#include namespace DB { diff --git a/src/Functions/URL/domainWithoutWWW.cpp b/src/Functions/URL/domainWithoutWWW.cpp index f6c8b5c84fc..436aad01b82 100644 --- a/src/Functions/URL/domainWithoutWWW.cpp +++ b/src/Functions/URL/domainWithoutWWW.cpp @@ -1,6 +1,6 @@ #include #include -#include "domain.h" +#include namespace DB { diff --git a/src/Functions/URL/firstSignificantSubdomain.cpp b/src/Functions/URL/firstSignificantSubdomain.cpp index b04f6d882ef..e929aefbc27 100644 --- a/src/Functions/URL/firstSignificantSubdomain.cpp +++ b/src/Functions/URL/firstSignificantSubdomain.cpp @@ -1,6 +1,6 @@ #include #include -#include "ExtractFirstSignificantSubdomain.h" +#include namespace DB diff --git a/src/Functions/URL/firstSignificantSubdomainCustom.cpp b/src/Functions/URL/firstSignificantSubdomainCustom.cpp index c07aa2b3ac8..95e5142667b 100644 --- a/src/Functions/URL/firstSignificantSubdomainCustom.cpp +++ b/src/Functions/URL/firstSignificantSubdomainCustom.cpp @@ -1,6 +1,6 @@ #include -#include "ExtractFirstSignificantSubdomain.h" -#include "FirstSignificantSubdomainCustomImpl.h" +#include +#include namespace DB diff --git a/src/Functions/URL/fragment.cpp b/src/Functions/URL/fragment.cpp index 262c1a4c7a6..66da5529d84 100644 --- a/src/Functions/URL/fragment.cpp +++ b/src/Functions/URL/fragment.cpp @@ -1,6 +1,6 @@ #include #include -#include "fragment.h" +#include namespace DB { diff --git a/src/Functions/URL/path.cpp b/src/Functions/URL/path.cpp index 8d609f43191..6de8e4fbf95 100644 --- a/src/Functions/URL/path.cpp +++ b/src/Functions/URL/path.cpp @@ -1,7 +1,7 @@ #include #include #include -#include "path.h" +#include #include diff --git a/src/Functions/URL/pathFull.cpp b/src/Functions/URL/pathFull.cpp index 9aacee21fed..deea617eb29 100644 --- a/src/Functions/URL/pathFull.cpp +++ b/src/Functions/URL/pathFull.cpp @@ -1,7 +1,7 @@ #include #include #include -#include "path.h" +#include #include namespace DB diff --git a/src/Functions/URL/port.cpp b/src/Functions/URL/port.cpp index c8f50f10a56..fac46281604 100644 --- a/src/Functions/URL/port.cpp +++ b/src/Functions/URL/port.cpp @@ -5,7 +5,7 @@ #include #include #include -#include "domain.h" +#include namespace DB diff --git a/src/Functions/URL/protocol.cpp b/src/Functions/URL/protocol.cpp index 6bed71207f6..1ac395dc554 100644 --- a/src/Functions/URL/protocol.cpp +++ b/src/Functions/URL/protocol.cpp @@ -1,6 +1,6 @@ #include #include -#include "protocol.h" +#include namespace DB diff --git a/src/Functions/URL/queryString.cpp b/src/Functions/URL/queryString.cpp index 7069ce46b38..f6aaf40fc96 100644 --- a/src/Functions/URL/queryString.cpp +++ b/src/Functions/URL/queryString.cpp @@ -1,6 +1,6 @@ #include #include -#include "queryString.h" +#include namespace DB { diff --git a/src/Functions/URL/queryStringAndFragment.cpp b/src/Functions/URL/queryStringAndFragment.cpp index 367a95acbdc..94f1dfa41e3 100644 --- a/src/Functions/URL/queryStringAndFragment.cpp +++ b/src/Functions/URL/queryStringAndFragment.cpp @@ -1,6 +1,6 @@ #include #include -#include "queryStringAndFragment.h" +#include namespace DB { diff --git a/src/Functions/URL/topLevelDomain.cpp b/src/Functions/URL/topLevelDomain.cpp index 25e9f383f60..b3e88832350 100644 --- a/src/Functions/URL/topLevelDomain.cpp +++ b/src/Functions/URL/topLevelDomain.cpp @@ -1,6 +1,6 @@ #include #include -#include "domain.h" +#include namespace DB { diff --git a/src/Functions/decodeHTMLComponent.cpp b/src/Functions/decodeHTMLComponent.cpp index 00a601b77a6..d36bee534a8 100644 --- a/src/Functions/decodeHTMLComponent.cpp +++ b/src/Functions/decodeHTMLComponent.cpp @@ -28,20 +28,20 @@ namespace const ColumnString::Chars & data, const ColumnString::Offsets & offsets, ColumnString::Chars & res_data, - ColumnString::Offsets & res_offsets) + ColumnString::Offsets & res_offsets, + size_t input_rows_count) { /// The size of result is always not more than the size of source. /// Because entities decodes to the shorter byte sequence. /// Example: &#xx... &#xx... will decode to UTF-8 byte sequence not longer than 4 bytes. res_data.resize(data.size()); - size_t size = offsets.size(); - res_offsets.resize(size); + res_offsets.resize(input_rows_count); size_t prev_offset = 0; size_t res_offset = 0; - for (size_t i = 0; i < size; ++i) + for (size_t i = 0; i < input_rows_count; ++i) { const char * src_data = reinterpret_cast(&data[prev_offset]); size_t src_size = offsets[i] - prev_offset; @@ -55,7 +55,7 @@ namespace res_data.resize(res_offset); } - [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &) + [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &, size_t) { throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, "Function decodeHTMLComponent cannot work with FixedString argument"); } @@ -64,7 +64,6 @@ namespace static const int max_legal_unicode_value = 0x10FFFF; static const int max_decimal_length_of_unicode_point = 7; /// 1114111 - static size_t execute(const char * src, size_t src_size, char * dst) { const char * src_pos = src; diff --git a/src/Functions/decodeXMLComponent.cpp b/src/Functions/decodeXMLComponent.cpp index cbbe46fcb8c..8743aa4759d 100644 --- a/src/Functions/decodeXMLComponent.cpp +++ b/src/Functions/decodeXMLComponent.cpp @@ -27,20 +27,20 @@ namespace const ColumnString::Chars & data, const ColumnString::Offsets & offsets, ColumnString::Chars & res_data, - ColumnString::Offsets & res_offsets) + ColumnString::Offsets & res_offsets, + size_t input_rows_count) { /// The size of result is always not more than the size of source. /// Because entities decodes to the shorter byte sequence. /// Example: &#xx... &#xx... will decode to UTF-8 byte sequence not longer than 4 bytes. res_data.resize(data.size()); - size_t size = offsets.size(); - res_offsets.resize(size); + res_offsets.resize(input_rows_count); size_t prev_offset = 0; size_t res_offset = 0; - for (size_t i = 0; i < size; ++i) + for (size_t i = 0; i < input_rows_count; ++i) { const char * src_data = reinterpret_cast(&data[prev_offset]); size_t src_size = offsets[i] - prev_offset; @@ -54,7 +54,7 @@ namespace res_data.resize(res_offset); } - [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &) + [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &, size_t) { throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, "Function decodeXMLComponent cannot work with FixedString argument"); } diff --git a/src/Functions/encodeXMLComponent.cpp b/src/Functions/encodeXMLComponent.cpp index 64d85ecaeb8..a22917838b7 100644 --- a/src/Functions/encodeXMLComponent.cpp +++ b/src/Functions/encodeXMLComponent.cpp @@ -25,17 +25,17 @@ namespace const ColumnString::Chars & data, const ColumnString::Offsets & offsets, ColumnString::Chars & res_data, - ColumnString::Offsets & res_offsets) + ColumnString::Offsets & res_offsets, + size_t input_rows_count) { /// 6 is the maximum size amplification (the maximum length of encoded entity: ") res_data.resize(data.size() * 6); - size_t size = offsets.size(); - res_offsets.resize(size); + res_offsets.resize(input_rows_count); size_t prev_offset = 0; size_t res_offset = 0; - for (size_t i = 0; i < size; ++i) + for (size_t i = 0; i < input_rows_count; ++i) { const char * src_data = reinterpret_cast(&data[prev_offset]); size_t src_size = offsets[i] - prev_offset; @@ -49,7 +49,7 @@ namespace res_data.resize(res_offset); } - [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &) + [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &, size_t) { throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, "Function encodeXML cannot work with FixedString argument"); } diff --git a/src/Functions/idna.cpp b/src/Functions/idna.cpp index 5a7ae3485ba..cf9e855c912 100644 --- a/src/Functions/idna.cpp +++ b/src/Functions/idna.cpp @@ -44,15 +44,15 @@ struct IdnaEncode const ColumnString::Chars & data, const ColumnString::Offsets & offsets, ColumnString::Chars & res_data, - ColumnString::Offsets & res_offsets) + ColumnString::Offsets & res_offsets, + size_t input_rows_count) { - const size_t rows = offsets.size(); res_data.reserve(data.size()); /// just a guess, assuming the input is all-ASCII - res_offsets.reserve(rows); + res_offsets.reserve(input_rows_count); size_t prev_offset = 0; std::string ascii; - for (size_t row = 0; row < rows; ++row) + for (size_t row = 0; row < input_rows_count; ++row) { const char * value = reinterpret_cast(&data[prev_offset]); const size_t value_length = offsets[row] - prev_offset - 1; @@ -85,7 +85,7 @@ struct IdnaEncode } } - [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &) + [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &, size_t) { throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Arguments of type FixedString are not allowed"); } @@ -99,15 +99,15 @@ struct IdnaDecode const ColumnString::Chars & data, const ColumnString::Offsets & offsets, ColumnString::Chars & res_data, - ColumnString::Offsets & res_offsets) + ColumnString::Offsets & res_offsets, + size_t input_rows_count) { - const size_t rows = offsets.size(); res_data.reserve(data.size()); /// just a guess, assuming the input is all-ASCII - res_offsets.reserve(rows); + res_offsets.reserve(input_rows_count); size_t prev_offset = 0; std::string unicode; - for (size_t row = 0; row < rows; ++row) + for (size_t row = 0; row < input_rows_count; ++row) { const char * ascii = reinterpret_cast(&data[prev_offset]); const size_t ascii_length = offsets[row] - prev_offset - 1; @@ -124,7 +124,7 @@ struct IdnaDecode } } - [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &) + [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &, size_t) { throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Arguments of type FixedString are not allowed"); } diff --git a/src/Functions/initcap.cpp b/src/Functions/initcap.cpp index 4661ce117c0..00414b22344 100644 --- a/src/Functions/initcap.cpp +++ b/src/Functions/initcap.cpp @@ -9,10 +9,12 @@ namespace struct InitcapImpl { - static void vector(const ColumnString::Chars & data, + static void vector( + const ColumnString::Chars & data, const ColumnString::Offsets & offsets, ColumnString::Chars & res_data, - ColumnString::Offsets & res_offsets) + ColumnString::Offsets & res_offsets, + size_t /*input_rows_count*/) { if (data.empty()) return; @@ -21,7 +23,7 @@ struct InitcapImpl array(data.data(), data.data() + data.size(), res_data.data()); } - static void vectorFixed(const ColumnString::Chars & data, size_t /*n*/, ColumnString::Chars & res_data) + static void vectorFixed(const ColumnString::Chars & data, size_t /*n*/, ColumnString::Chars & res_data, size_t) { res_data.resize(data.size()); array(data.data(), data.data() + data.size(), res_data.data()); diff --git a/src/Functions/initcapUTF8.cpp b/src/Functions/initcapUTF8.cpp index 076dcff6622..282d846094e 100644 --- a/src/Functions/initcapUTF8.cpp +++ b/src/Functions/initcapUTF8.cpp @@ -22,7 +22,8 @@ struct InitcapUTF8Impl const ColumnString::Chars & data, const ColumnString::Offsets & offsets, ColumnString::Chars & res_data, - ColumnString::Offsets & res_offsets) + ColumnString::Offsets & res_offsets, + size_t /*input_rows_count*/) { if (data.empty()) return; @@ -31,7 +32,7 @@ struct InitcapUTF8Impl array(data.data(), data.data() + data.size(), offsets, res_data.data()); } - [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &) + [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &, size_t) { throw Exception(ErrorCodes::BAD_ARGUMENTS, "Function initcapUTF8 cannot work with FixedString argument"); } diff --git a/src/Functions/normalizeQuery.cpp b/src/Functions/normalizeQuery.cpp index ad9a8903733..1a78bce7d29 100644 --- a/src/Functions/normalizeQuery.cpp +++ b/src/Functions/normalizeQuery.cpp @@ -19,17 +19,19 @@ template struct Impl { static constexpr auto name = keep_names ? "normalizeQueryKeepNames" : "normalizeQuery"; - static void vector(const ColumnString::Chars & data, + + static void vector( + const ColumnString::Chars & data, const ColumnString::Offsets & offsets, ColumnString::Chars & res_data, - ColumnString::Offsets & res_offsets) + ColumnString::Offsets & res_offsets, + size_t input_rows_count) { - size_t size = offsets.size(); - res_offsets.resize(size); + res_offsets.resize(input_rows_count); res_data.reserve(data.size()); ColumnString::Offset prev_src_offset = 0; - for (size_t i = 0; i < size; ++i) + for (size_t i = 0; i < input_rows_count; ++i) { ColumnString::Offset curr_src_offset = offsets[i]; @@ -43,7 +45,7 @@ struct Impl } } - [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &) + [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &, size_t) { throw Exception(ErrorCodes::ILLEGAL_COLUMN, "Cannot apply function normalizeQuery to fixed string."); } diff --git a/src/Functions/normalizeString.cpp b/src/Functions/normalizeString.cpp index 92411490eaa..c56508ced0e 100644 --- a/src/Functions/normalizeString.cpp +++ b/src/Functions/normalizeString.cpp @@ -84,7 +84,8 @@ struct NormalizeUTF8Impl static void vector(const ColumnString::Chars & data, const ColumnString::Offsets & offsets, ColumnString::Chars & res_data, - ColumnString::Offsets & res_offsets) + ColumnString::Offsets & res_offsets, + size_t input_rows_count) { UErrorCode err = U_ZERO_ERROR; @@ -92,8 +93,7 @@ struct NormalizeUTF8Impl if (U_FAILURE(err)) throw Exception(ErrorCodes::CANNOT_NORMALIZE_STRING, "Normalization failed (getNormalizer): {}", u_errorName(err)); - size_t size = offsets.size(); - res_offsets.resize(size); + res_offsets.resize(input_rows_count); res_data.reserve(data.size() * 2); @@ -103,7 +103,7 @@ struct NormalizeUTF8Impl PODArray from_uchars; PODArray to_uchars; - for (size_t i = 0; i < size; ++i) + for (size_t i = 0; i < input_rows_count; ++i) { size_t from_size = offsets[i] - current_from_offset - 1; @@ -157,7 +157,7 @@ struct NormalizeUTF8Impl res_data.resize(current_to_offset); } - [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &) + [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &, size_t) { throw Exception(ErrorCodes::ILLEGAL_COLUMN, "Cannot apply function normalizeUTF8 to fixed string."); } diff --git a/src/Functions/normalizedQueryHash.cpp b/src/Functions/normalizedQueryHash.cpp index 63218f28af5..3dbe3ff9847 100644 --- a/src/Functions/normalizedQueryHash.cpp +++ b/src/Functions/normalizedQueryHash.cpp @@ -27,13 +27,13 @@ struct Impl static void vector( const ColumnString::Chars & data, const ColumnString::Offsets & offsets, - PaddedPODArray & res_data) + PaddedPODArray & res_data, + size_t input_rows_count) { - size_t size = offsets.size(); - res_data.resize(size); + res_data.resize(input_rows_count); ColumnString::Offset prev_src_offset = 0; - for (size_t i = 0; i < size; ++i) + for (size_t i = 0; i < input_rows_count; ++i) { ColumnString::Offset curr_src_offset = offsets[i]; res_data[i] = normalizedQueryHash( @@ -77,15 +77,15 @@ public: bool isSuitableForShortCircuitArgumentsExecution(const DataTypesWithConstInfo & /*arguments*/) const override { return true; } - ColumnPtr executeImpl(const ColumnsWithTypeAndName & arguments, const DataTypePtr &, size_t /*input_rows_count*/) const override + ColumnPtr executeImpl(const ColumnsWithTypeAndName & arguments, const DataTypePtr &, size_t input_rows_count) const override { const ColumnPtr column = arguments[0].column; if (const ColumnString * col = checkAndGetColumn(column.get())) { auto col_res = ColumnUInt64::create(); typename ColumnUInt64::Container & vec_res = col_res->getData(); - vec_res.resize(col->size()); - Impl::vector(col->getChars(), col->getOffsets(), vec_res); + vec_res.resize(input_rows_count); + Impl::vector(col->getChars(), col->getOffsets(), vec_res, input_rows_count); return col_res; } else diff --git a/src/Functions/pointInEllipses.cpp b/src/Functions/pointInEllipses.cpp index 2147428cee3..ac7a8cc4204 100644 --- a/src/Functions/pointInEllipses.cpp +++ b/src/Functions/pointInEllipses.cpp @@ -91,8 +91,6 @@ private: ColumnPtr executeImpl(const ColumnsWithTypeAndName & arguments, const DataTypePtr &, size_t input_rows_count) const override { - const auto size = input_rows_count; - /// Prepare array of ellipses. size_t ellipses_count = (arguments.size() - 2) / 4; std::vector ellipses(ellipses_count); @@ -141,13 +139,11 @@ private: auto dst = ColumnVector::create(); auto & dst_data = dst->getData(); - dst_data.resize(size); + dst_data.resize(input_rows_count); size_t start_index = 0; - for (const auto row : collections::range(0, size)) - { + for (size_t row = 0; row < input_rows_count; ++row) dst_data[row] = isPointInEllipses(col_vec_x->getData()[row], col_vec_y->getData()[row], ellipses.data(), ellipses_count, start_index); - } return dst; } @@ -157,7 +153,7 @@ private: const auto * col_const_y = assert_cast (col_y); size_t start_index = 0; UInt8 res = isPointInEllipses(col_const_x->getValue(), col_const_y->getValue(), ellipses.data(), ellipses_count, start_index); - return DataTypeUInt8().createColumnConst(size, res); + return DataTypeUInt8().createColumnConst(input_rows_count, res); } else { diff --git a/src/Functions/punycode.cpp b/src/Functions/punycode.cpp index 8004e3731b5..ec1fcfd0a70 100644 --- a/src/Functions/punycode.cpp +++ b/src/Functions/punycode.cpp @@ -6,11 +6,11 @@ #include #include -# pragma clang diagnostic push -# pragma clang diagnostic ignored "-Wnewline-eof" -# include -# include -# pragma clang diagnostic pop +#pragma clang diagnostic push +#pragma clang diagnostic ignored "-Wnewline-eof" +#include +#include +#pragma clang diagnostic pop namespace DB { @@ -38,16 +38,16 @@ struct PunycodeEncode const ColumnString::Chars & data, const ColumnString::Offsets & offsets, ColumnString::Chars & res_data, - ColumnString::Offsets & res_offsets) + ColumnString::Offsets & res_offsets, + size_t input_rows_count) { - const size_t rows = offsets.size(); res_data.reserve(data.size()); /// just a guess, assuming the input is all-ASCII - res_offsets.reserve(rows); + res_offsets.reserve(input_rows_count); size_t prev_offset = 0; std::u32string value_utf32; std::string value_puny; - for (size_t row = 0; row < rows; ++row) + for (size_t row = 0; row < input_rows_count; ++row) { const char * value = reinterpret_cast(&data[prev_offset]); const size_t value_length = offsets[row] - prev_offset - 1; @@ -72,7 +72,7 @@ struct PunycodeEncode } } - [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &) + [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &, size_t) { throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Arguments of type FixedString are not allowed"); } @@ -86,16 +86,16 @@ struct PunycodeDecode const ColumnString::Chars & data, const ColumnString::Offsets & offsets, ColumnString::Chars & res_data, - ColumnString::Offsets & res_offsets) + ColumnString::Offsets & res_offsets, + size_t input_rows_count) { - const size_t rows = offsets.size(); res_data.reserve(data.size()); /// just a guess, assuming the input is all-ASCII - res_offsets.reserve(rows); + res_offsets.reserve(input_rows_count); size_t prev_offset = 0; std::u32string value_utf32; std::string value_utf8; - for (size_t row = 0; row < rows; ++row) + for (size_t row = 0; row < input_rows_count; ++row) { const char * value = reinterpret_cast(&data[prev_offset]); const size_t value_length = offsets[row] - prev_offset - 1; @@ -129,7 +129,7 @@ struct PunycodeDecode } } - [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &) + [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &, size_t) { throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Arguments of type FixedString are not allowed"); } diff --git a/src/Functions/reverse.cpp b/src/Functions/reverse.cpp index d23e48b8d42..94b6634ffbd 100644 --- a/src/Functions/reverse.cpp +++ b/src/Functions/reverse.cpp @@ -55,19 +55,19 @@ public: bool useDefaultImplementationForConstants() const override { return true; } - ColumnPtr executeImpl(const ColumnsWithTypeAndName & arguments, const DataTypePtr &, size_t) const override + ColumnPtr executeImpl(const ColumnsWithTypeAndName & arguments, const DataTypePtr &, size_t input_rows_count) const override { const ColumnPtr column = arguments[0].column; if (const ColumnString * col = checkAndGetColumn(column.get())) { auto col_res = ColumnString::create(); - ReverseImpl::vector(col->getChars(), col->getOffsets(), col_res->getChars(), col_res->getOffsets()); + ReverseImpl::vector(col->getChars(), col->getOffsets(), col_res->getChars(), col_res->getOffsets(), input_rows_count); return col_res; } else if (const ColumnFixedString * col_fixed = checkAndGetColumn(column.get())) { auto col_res = ColumnFixedString::create(col_fixed->getN()); - ReverseImpl::vectorFixed(col_fixed->getChars(), col_fixed->getN(), col_res->getChars()); + ReverseImpl::vectorFixed(col_fixed->getChars(), col_fixed->getN(), col_res->getChars(), input_rows_count); return col_res; } else diff --git a/src/Functions/reverse.h b/src/Functions/reverse.h index 5f999af4297..7c18d72769a 100644 --- a/src/Functions/reverse.h +++ b/src/Functions/reverse.h @@ -9,17 +9,18 @@ namespace DB */ struct ReverseImpl { - static void vector(const ColumnString::Chars & data, + static void vector( + const ColumnString::Chars & data, const ColumnString::Offsets & offsets, ColumnString::Chars & res_data, - ColumnString::Offsets & res_offsets) + ColumnString::Offsets & res_offsets, + size_t input_rows_count) { res_data.resize_exact(data.size()); res_offsets.assign(offsets); - size_t size = offsets.size(); ColumnString::Offset prev_offset = 0; - for (size_t i = 0; i < size; ++i) + for (size_t i = 0; i < input_rows_count; ++i) { for (size_t j = prev_offset; j < offsets[i] - 1; ++j) res_data[j] = data[offsets[i] + prev_offset - 2 - j]; @@ -28,12 +29,15 @@ struct ReverseImpl } } - static void vectorFixed(const ColumnString::Chars & data, size_t n, ColumnString::Chars & res_data) + static void vectorFixed( + const ColumnString::Chars & data, + size_t n, + ColumnString::Chars & res_data, + size_t input_rows_count) { res_data.resize_exact(data.size()); - size_t size = data.size() / n; - for (size_t i = 0; i < size; ++i) + for (size_t i = 0; i < input_rows_count; ++i) for (size_t j = i * n; j < (i + 1) * n; ++j) res_data[j] = data[(i * 2 + 1) * n - j - 1]; } diff --git a/src/Functions/reverseUTF8.cpp b/src/Functions/reverseUTF8.cpp index 1aee349fa8d..ca57d946b19 100644 --- a/src/Functions/reverseUTF8.cpp +++ b/src/Functions/reverseUTF8.cpp @@ -23,25 +23,25 @@ namespace */ struct ReverseUTF8Impl { - static void vector(const ColumnString::Chars & data, + static void vector( + const ColumnString::Chars & data, const ColumnString::Offsets & offsets, ColumnString::Chars & res_data, - ColumnString::Offsets & res_offsets) + ColumnString::Offsets & res_offsets, + size_t input_rows_count) { bool all_ascii = isAllASCII(data.data(), data.size()); if (all_ascii) { - ReverseImpl::vector(data, offsets, res_data, res_offsets); + ReverseImpl::vector(data, offsets, res_data, res_offsets, input_rows_count); return; } res_data.resize(data.size()); res_offsets.assign(offsets); - size_t size = offsets.size(); - ColumnString::Offset prev_offset = 0; - for (size_t i = 0; i < size; ++i) + for (size_t i = 0; i < input_rows_count; ++i) { ColumnString::Offset j = prev_offset; while (j < offsets[i] - 1) @@ -73,7 +73,7 @@ struct ReverseUTF8Impl } } - [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &) + [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &, size_t) { throw Exception(ErrorCodes::ILLEGAL_COLUMN, "Cannot apply function reverseUTF8 to fixed string."); } diff --git a/src/Functions/soundex.cpp b/src/Functions/soundex.cpp index fcf1523d1a3..e8bd1b664e3 100644 --- a/src/Functions/soundex.cpp +++ b/src/Functions/soundex.cpp @@ -79,14 +79,14 @@ struct SoundexImpl const ColumnString::Chars & data, const ColumnString::Offsets & offsets, ColumnString::Chars & res_data, - ColumnString::Offsets & res_offsets) + ColumnString::Offsets & res_offsets, + size_t input_rows_count) { - const size_t size = offsets.size(); - res_data.resize(size * (length + 1)); - res_offsets.resize(size); + res_data.resize(input_rows_count * (length + 1)); + res_offsets.resize(input_rows_count); size_t prev_offset = 0; - for (size_t i = 0; i < size; ++i) + for (size_t i = 0; i < input_rows_count; ++i) { const char * value = reinterpret_cast(&data[prev_offset]); const size_t value_length = offsets[i] - prev_offset - 1; @@ -98,7 +98,7 @@ struct SoundexImpl } } - [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &) + [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &, size_t) { throw Exception(ErrorCodes::ILLEGAL_COLUMN, "Column of type FixedString is not supported by soundex function"); } diff --git a/src/Functions/toValidUTF8.cpp b/src/Functions/toValidUTF8.cpp index 41d29d9c494..376732256b0 100644 --- a/src/Functions/toValidUTF8.cpp +++ b/src/Functions/toValidUTF8.cpp @@ -128,16 +128,16 @@ struct ToValidUTF8Impl const ColumnString::Chars & data, const ColumnString::Offsets & offsets, ColumnString::Chars & res_data, - ColumnString::Offsets & res_offsets) + ColumnString::Offsets & res_offsets, + size_t input_rows_count) { - const size_t offsets_size = offsets.size(); /// It can be larger than that, but we believe it is unlikely to happen. res_data.resize(data.size()); - res_offsets.resize(offsets_size); + res_offsets.resize(input_rows_count); size_t prev_offset = 0; WriteBufferFromVector write_buffer(res_data); - for (size_t i = 0; i < offsets_size; ++i) + for (size_t i = 0; i < input_rows_count; ++i) { const char * haystack_data = reinterpret_cast(&data[prev_offset]); const size_t haystack_size = offsets[i] - prev_offset - 1; @@ -149,7 +149,7 @@ struct ToValidUTF8Impl write_buffer.finalize(); } - [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &) + [[noreturn]] static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &, size_t) { throw Exception(ErrorCodes::ILLEGAL_COLUMN, "Column of type FixedString is not supported by toValidUTF8 function"); } diff --git a/src/Functions/trim.cpp b/src/Functions/trim.cpp index 1f0011b8e99..5703e871423 100644 --- a/src/Functions/trim.cpp +++ b/src/Functions/trim.cpp @@ -43,10 +43,10 @@ public: const ColumnString::Chars & data, const ColumnString::Offsets & offsets, ColumnString::Chars & res_data, - ColumnString::Offsets & res_offsets) + ColumnString::Offsets & res_offsets, + size_t input_rows_count) { - size_t size = offsets.size(); - res_offsets.resize_exact(size); + res_offsets.resize_exact(input_rows_count); res_data.reserve_exact(data.size()); size_t prev_offset = 0; @@ -55,7 +55,7 @@ public: const UInt8 * start; size_t length; - for (size_t i = 0; i < size; ++i) + for (size_t i = 0; i < input_rows_count; ++i) { execute(reinterpret_cast(&data[prev_offset]), offsets[i] - prev_offset - 1, start, length); @@ -69,7 +69,7 @@ public: } } - static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &) + static void vectorFixed(const ColumnString::Chars &, size_t, ColumnString::Chars &, size_t) { throw Exception(ErrorCodes::ILLEGAL_TYPE_OF_ARGUMENT, "Functions trimLeft, trimRight and trimBoth cannot work with FixedString argument"); }