From 17d6a993fc17d533460c5f40f3908c708e057c18 Mon Sep 17 00:00:00 2001 From: Daniel Baumann Date: Thu, 23 May 2024 18:45:17 +0200 Subject: Merging upstream version 18.2.3. Signed-off-by: Daniel Baumann --- src/s3select/include/s3select.h | 271 ++++++++++++++++++++++------ src/s3select/include/s3select_csv_parser.h | 67 ++++++- src/s3select/include/s3select_functions.h | 110 +++++++++-- src/s3select/include/s3select_json_parser.h | 1 - src/s3select/include/s3select_oper.h | 35 +++- 5 files changed, 403 insertions(+), 81 deletions(-) (limited to 'src/s3select/include') diff --git a/src/s3select/include/s3select.h b/src/s3select/include/s3select.h index 3ac111351..667c92ba9 100644 --- a/src/s3select/include/s3select.h +++ b/src/s3select/include/s3select.h @@ -18,6 +18,7 @@ #include #include #include +#include #define _DEBUG_TERM {string token(a,b);std::cout << __FUNCTION__ << token << std::endl;} @@ -1293,11 +1294,11 @@ void push_logical_operator::builder(s3select* self, const char* a, const char* b std::string token(a, b); logical_operand::oplog_t l = logical_operand::oplog_t::NA; - if (token == "and") + if (boost::iequals(token,"and")) { l = logical_operand::oplog_t::AND; } - else if (token == "or") + else if (boost::iequals(token,"or")) { l = logical_operand::oplog_t::OR; } @@ -1633,8 +1634,10 @@ void push_like_predicate_escape::builder(s3select* self, const char* a, const ch void push_is_null_predicate::builder(s3select* self, const char* a, const char* b) const { - //expression is null, is not null + //expression could be is null OR is not null std::string token(a, b); + //to_lower enable case insensitive + boost::algorithm::to_lower(token); bool is_null = true; for(size_t i=0;i comment_chars; std::vector trim_chars; + std::string schema; - s3select_csv_definitions():row_delimiter('\n'), column_delimiter(','), output_row_delimiter('\n'), output_column_delimiter(','), escape_char('\\'), output_escape_char('\\'), output_quot_char('"'), quot_char('"'), use_header_info(false), ignore_header_info(false), quote_fields_always(false), quote_fields_asneeded(false), redundant_column(false), comment_empty_lines(false) {} + s3select_csv_definitions():row_delimiter('\n'), column_delimiter(','), output_row_delimiter('\n'), output_column_delimiter(','), escape_char('\\'), output_escape_char('\\'), output_quot_char('"'), quot_char('"'), use_header_info(false), ignore_header_info(false), quote_fields_always(false), quote_fields_asneeded(false), redundant_column(false), comment_empty_lines(false), output_json_format(false) {} }; @@ -2172,6 +2177,8 @@ protected: scratch_area* m_sa; std::string m_obj_name; bool m_aggr_flow = false; //TODO once per query + bool is_star = false; + bool is_json = false; bool m_is_to_aggregate; std::vector m_projections; base_statement* m_where_clause; @@ -2182,9 +2189,11 @@ protected: unsigned long m_processed_rows; size_t m_returned_bytes_size; std::function fp_ext_debug_mesg;//dispache debug message into external system + std::vector m_projection_keys{}; public: s3select_csv_definitions m_csv_defintion;//TODO add method for modify + std::string m_error_description; enum class Status { END_OF_STREAM, @@ -2196,6 +2205,16 @@ public: Status m_sql_processing_status; + void set_processing_time_error() + { + m_sql_processing_status = Status::SQL_ERROR; + } + + bool is_processing_time_error() + { + return m_sql_processing_status == Status::SQL_ERROR; + } + Status get_sql_processing_status() { return m_sql_processing_status; @@ -2206,6 +2225,60 @@ public: return m_sql_processing_status == Status::LIMIT_REACHED; } + void set_star_true() { + is_star = true; + } + + void set_projection_keys(std::vector m_projections) + { + std::vector alias_values{}; + std::unordered_set alias_projection_keys{}; + bool is_output_json_format = m_csv_defintion.output_json_format; + + for (auto& a : *m_s3_select->get_aliases()->get()) + { + alias_values.push_back(a.first); + alias_projection_keys.insert(a.second); + } + + size_t m_alias_index = 0; + int index_json_projection = 0; + is_json = m_s3_select->is_json_query(); + + for (auto& p : m_projections) + { + if(p->is_statement_contain_star_operation()) + { + set_star_true(); + } + p->traverse_and_apply(m_sa, m_s3_select->get_aliases(), m_s3_select->is_json_query()); + + std::string key_from_projection{}; + if(p->is_column()){ + key_from_projection = p->get_key_from_projection(); + } + + if(alias_projection_keys.count(p) == 0 && p->is_column()) { + m_projection_keys.push_back(key_from_projection); + } else if(alias_projection_keys.count(p) > 0 && p->is_column()) { + m_projection_keys.push_back(alias_values[m_alias_index++]); + } else if(!p->is_column() && is_output_json_format && alias_projection_keys.count(p) > 0 ) { + m_projection_keys.push_back(alias_values[m_alias_index++]); + } else if(!p->is_column() && is_output_json_format && alias_projection_keys.count(p) == 0) { + std::string index_json = "_" + std::to_string(++index_json_projection); + m_projection_keys.push_back(index_json); + } + } + + if(m_s3_select->is_json_query()) { + for(auto& k: m_projection_keys) { + size_t lastDotPosition = k.find_last_of('.'); + std::string extractedPart = k.substr(lastDotPosition + 1); + k = extractedPart; + } + } + } + void set_base_defintions(s3select* m) { if(m_s3_select || !m) @@ -2225,10 +2298,8 @@ public: m_where_clause->traverse_and_apply(m_sa, m_s3_select->get_aliases(), m_s3_select->is_json_query()); } - for (auto& p : m_projections) - { - p->traverse_and_apply(m_sa, m_s3_select->get_aliases(), m_s3_select->is_json_query()); - } + set_projection_keys(m_projections); + m_is_to_aggregate = true;//TODO not correct. should be set upon end-of-stream m_aggr_flow = m_s3_select->is_aggregate_query(); @@ -2269,45 +2340,95 @@ public: return m_returned_bytes_size; } - void result_values_to_string(multi_values& projections_resuls, std::string& result) + void json_result_format(multi_values& projections_results, std::string& result, std::string& output_delimiter) { - size_t i = 0; + result += "{"; + int j = 0; + for (size_t i = 0; i < projections_results.values.size(); ++i) + { + auto& res = projections_results.values[i]; + std::string label = "_"; + label += std::to_string(i + 1); + + if (i > 0) { + result += output_delimiter; + } + + if(!is_star) { + result += "\"" + m_projection_keys[j] + "\":"; + } else if(is_star && !is_json) { + result += "\"" + label + "\":"; + } + + result.append(res->to_string()); + m_returned_bytes_size += strlen(res->to_string()); + ++j; + } + result += "}"; + + } + + + void result_values_to_string(multi_values& projections_resuls, std::string& result) +{ std::string output_delimiter(1,m_csv_defintion.output_column_delimiter); std::string output_row_delimiter(1,m_csv_defintion.output_row_delimiter); + if(m_csv_defintion.output_json_format && projections_resuls.values.size()) { + json_result_format(projections_resuls, result, output_delimiter); + result.append(output_row_delimiter); + return; + } + + size_t i = 0; for(auto& res : projections_resuls.values) { + + std::string column_result; + + try{ + column_result = res->to_string(); + } + catch(std::exception& e) + { + column_result = "{failed to compute projection: " + std::string(e.what()) + "}"; + m_error_description = column_result; + set_processing_time_error(); + } + + if(fp_ext_debug_mesg) - fp_ext_debug_mesg( res->to_string() ); + fp_ext_debug_mesg(column_result.data()); if (m_csv_defintion.quote_fields_always) { std::ostringstream quoted_result; - quoted_result << std::quoted(res->to_string(),m_csv_defintion.output_quot_char, m_csv_defintion.escape_char); + quoted_result << std::quoted(column_result,m_csv_defintion.output_quot_char, m_csv_defintion.escape_char); result.append(quoted_result.str()); + m_returned_bytes_size += quoted_result.str().size(); - }//TODO to add asneeded + }//TODO to add asneeded else { - result.append(res->to_string()); - m_returned_bytes_size += strlen(res->to_string()); + result.append(column_result); + m_returned_bytes_size += column_result.size(); + } - if(!m_csv_defintion.redundant_column) { - if(++i < projections_resuls.values.size()) { - result.append(output_delimiter); - m_returned_bytes_size += output_delimiter.size(); - } - } - else { - result.append(output_delimiter); - m_returned_bytes_size += output_delimiter.size(); - } - } - if(!m_aggr_flow){ - result.append(output_row_delimiter); + if(!m_csv_defintion.redundant_column) { + if(++i < projections_resuls.values.size()) { + result.append(output_delimiter); + m_returned_bytes_size += output_delimiter.size(); + } + } else { + result.append(output_delimiter); m_returned_bytes_size += output_delimiter.size(); + } } - } + if(!m_aggr_flow) { + result.append(output_row_delimiter); + m_returned_bytes_size += output_delimiter.size(); + } +} Status getMatchRow( std::string& result) { @@ -2336,7 +2457,7 @@ public: } result_values_to_string(projections_resuls,result); - return m_sql_processing_status = Status::END_OF_STREAM; + return is_processing_time_error() ? (m_sql_processing_status = Status::SQL_ERROR) : (m_sql_processing_status = Status::END_OF_STREAM); } m_processed_rows++; @@ -2368,9 +2489,9 @@ public: i->set_last_call(); i->set_skip_non_aggregate(false);//projection column is set to be runnable projections_resuls.push_value( &(i->eval()) ); - } + } result_values_to_string(projections_resuls,result); - return m_sql_processing_status = Status::LIMIT_REACHED; + return is_processing_time_error() ? (m_sql_processing_status = Status::SQL_ERROR) : (m_sql_processing_status = Status::LIMIT_REACHED); } } while (multiple_row_processing()); @@ -2394,6 +2515,7 @@ public: { a.second->invalidate_cache_result(); } + } while (multiple_row_processing() && m_where_clause && !(where_clause_result = m_where_clause->eval().is_true()) && !(m_is_limit_on && m_processed_rows == m_limit)); @@ -2421,12 +2543,18 @@ public: for (auto& i : m_projections) { projections_resuls.push_value( &(i->eval()) ); - } - result_values_to_string(projections_resuls,result); + } + result_values_to_string(projections_resuls,result); + if(m_sql_processing_status == Status::SQL_ERROR) + { + return m_sql_processing_status; + } } - } - return is_end_of_stream() ? (m_sql_processing_status = Status::END_OF_STREAM) : (m_sql_processing_status = Status::NORMAL_EXIT); + + return is_processing_time_error() ? (m_sql_processing_status = Status::SQL_ERROR) : + (is_end_of_stream() ? (m_sql_processing_status = Status::END_OF_STREAM) : (m_sql_processing_status = Status::NORMAL_EXIT)); + }//getMatchRow @@ -2477,14 +2605,12 @@ public: { //return; } - - set_base_defintions(s3_query); m_csv_defintion = csv; + set_base_defintions(s3_query); } private: bool m_skip_last_line; - std::string m_error_description; char* m_stream; char* m_end_stream; std::vector m_row_tokens; @@ -2604,26 +2730,36 @@ public: m_error_description = "escaped_char_missing failure while csv parsing"; return -1; } - catch(io::error::escaped_string_not_closed& err) + catch(io::error::escaped_string_not_closed& err) { m_error_description = "escaped_string_not_closed failure while csv parsing"; return -1; } - catch(io::error::line_length_limit_exceeded& err) + catch(io::error::line_length_limit_exceeded& err) { m_error_description = "line_length_limit_exceeded failure while csv parsing"; return -1; } - catch(io::error::with_file_name& err) + catch(io::error::missmatch_of_begin_end& err) { - m_error_description = "with_file_name failure while csv parsing"; + m_error_description = "missmatch_of_begin_end failure while csv parsing" + std::string(err.what()); return -1; } - catch(io::error::with_file_line& err) + catch(io::error::missmatch_end& err) { - m_error_description = "with_file_line failure while csv parsing"; + m_error_description = "missmatch_end failure while csv parsing" + std::string(err.what()); return -1; } + catch(io::error::with_file_name& err) + { + m_error_description = "with_file_name failure while csv parsing"; + return -1; + } + catch(std::exception& e) + { + m_error_description = "error while processing CSV object : " + std::string(e.what()); + return -1; + } return status; } @@ -2634,7 +2770,7 @@ private: //purpose: the CSV data is "streaming", it may "cut" rows in the middle, in that case the "broken-line" is stores //for later, upon next chunk of data is streaming, the stored-line is merge with current broken-line, and processed. std::string tmp_buff; - + int status = 0; m_processed_bytes += stream_length; m_skip_first_line = false; @@ -2648,6 +2784,22 @@ private: p_obj_chunk++; } + if(*p_obj_chunk != m_csv_defintion.row_delimiter) + {// previous row can not be completed with current chunk + if(fp_ext_debug_mesg) + { + std::string err_mesg = "** the stream chunk is too small for processing(saved for later) **"; + fp_ext_debug_mesg(err_mesg.c_str()); + } + //copy the part to be processed later + tmp_buff.assign((char*)csv_stream, (char*)csv_stream + (p_obj_chunk - csv_stream)); + //saved for later processing + m_last_line.append(tmp_buff); + m_previous_line = true;//it means to skip last line + //skip processing since the row tail is missing. + return 0; + } + tmp_buff.assign((char*)csv_stream, (char*)csv_stream + (p_obj_chunk - csv_stream)); merge_line = m_last_line + tmp_buff + m_csv_defintion.row_delimiter; m_previous_line = false; @@ -2655,7 +2807,7 @@ private: m_skip_x_first_bytes = tmp_buff.size()+1; //processing the merged row (previous broken row) - run_s3select_on_object(result, merge_line.c_str(), merge_line.length(), false, false, false); + status = run_s3select_on_object(result, merge_line.c_str(), merge_line.length(), false, false, false); } if (stream_length && csv_stream[stream_length - 1] != m_csv_defintion.row_delimiter) @@ -2676,7 +2828,8 @@ private: stream_length -= (m_last_line.length()); } - return run_s3select_on_object(result, csv_stream, stream_length, m_skip_first_line, m_previous_line, (m_processed_bytes >= obj_size)); + status = run_s3select_on_object(result, csv_stream, stream_length, m_skip_first_line, m_previous_line, (m_processed_bytes >= obj_size)); + return status; } public: @@ -2696,6 +2849,11 @@ public: m_skip_x_first_bytes=0; } + if(m_stream>m_end_stream) + { + throw base_s3select_exception(std::string("** m_stream > m_end_stream **") + + std::to_string( (m_stream - m_end_stream) ) ,base_s3select_exception::s3select_exp_en_t::FATAL); + } CSVParser _csv_parser("csv", m_stream, m_end_stream); csv_parser = &_csv_parser; csv_parser->set_csv_def( m_csv_defintion.row_delimiter, @@ -2745,6 +2903,10 @@ public: { break;//user should request for sql_processing_status } + if(m_sql_processing_status == Status::SQL_ERROR) + { + return -1; + } } while (true); @@ -2764,7 +2926,6 @@ class parquet_object : public base_s3object { private: - std::string m_error_description; parquet_file_parser* object_reader; parquet_file_parser::column_pos_t m_where_clause_columns; parquet_file_parser::column_pos_t m_projections_columns; @@ -2948,11 +3109,13 @@ private: std::string* m_s3select_result = nullptr; size_t m_row_count; bool star_operation_ind; - std::string m_error_description; bool m_init_json_processor_ind; public: + class csv_definitions : public s3select_csv_definitions + {}; + void init_json_processor(s3select* query) { if(m_init_json_processor_ind) @@ -2997,6 +3160,7 @@ public: if(p->is_statement_contain_star_operation()) { star_operation_ind=true; + set_star_true(); break; } } @@ -3095,7 +3259,7 @@ private: public: - int run_s3select_on_stream(std::string& result, const char* json_stream, size_t stream_length, size_t obj_size) + int run_s3select_on_stream(std::string& result, const char* json_stream, size_t stream_length, size_t obj_size, bool json_format = false) { int status=0; m_processed_bytes += stream_length; @@ -3134,8 +3298,9 @@ public: return status; } - void set_json_query(s3select* s3_query) + void set_json_query(s3select* s3_query, csv_definitions csv) { + m_csv_defintion = csv; set_base_defintions(s3_query); init_json_processor(s3_query); } diff --git a/src/s3select/include/s3select_csv_parser.h b/src/s3select/include/s3select_csv_parser.h index dab2e4efa..28e8117a6 100644 --- a/src/s3select/include/s3select_csv_parser.h +++ b/src/s3select/include/s3select_csv_parser.h @@ -13,6 +13,53 @@ namespace io{ , file_line, file_name); } }; + + struct missmatch_of_begin_end : + base, + with_file_name, + with_file_line{ + int begin=-1,end=-1; + void set_begin_end(int b,int e){ + begin=b; + end=e; + } + + void format_error_message()const override{ + std::snprintf(error_message_buffer, sizeof(error_message_buffer), + "***missmatch_of_begin_end*** Line number %d in file \"%s\" begin{%d} > end{%d}" + ,file_line, file_name,begin,end); + } + }; + + struct missmatch_end : + base, + with_file_name, + with_file_line{ + int end=-1; + int block_size=-1; + void set_end_block(int e,int b){ + end = e; + block_size = b; + } + void format_error_message()const override{ + std::snprintf(error_message_buffer, sizeof(error_message_buffer), + "***missmatch_end*** Line number %d in file \"%s\" end{%d} block{%d}" + ,file_line, file_name, end, block_size); + } + }; + + + struct line_is_null : + base, + with_file_name, + with_file_line{ + void format_error_message()const override{ + std::snprintf(error_message_buffer, sizeof(error_message_buffer), + "***line is NULL*** Line number %d in file \"%s\"" + ,file_line, file_name); + } + }; + } namespace detail{ @@ -133,7 +180,11 @@ namespace io{ void chop_next_column(char*&line, char*&col_begin, char*&col_end, char& col_delimiter, char& quote, char& escape_char) { - assert(line != nullptr); + if(line == NULL) + { + io::error::line_is_null err; + throw err; + } col_begin = line; // the col_begin + (... - col_begin) removes the constness @@ -312,8 +363,18 @@ class CSVParser ++file_line; - assert(data_begin < data_end); - assert(data_end <= block_len*2); + if(data_begin > data_end) + { + io::error::missmatch_of_begin_end err; + err.set_begin_end(data_begin,data_end); + throw err; + } + if(data_end > block_len*2) + { + io::error::missmatch_end err; + err.set_end_block(data_end,block_len*2); + throw err; + } if(data_begin >= block_len) { diff --git a/src/s3select/include/s3select_functions.h b/src/s3select/include/s3select_functions.h index 8c507fca1..4d88d772e 100644 --- a/src/s3select/include/s3select_functions.h +++ b/src/s3select/include/s3select_functions.h @@ -350,6 +350,7 @@ private: s3select_functions* m_s3select_functions; variable m_result; bool m_is_aggregate_function; + value eval_result; void _resolve_name() { @@ -437,14 +438,28 @@ public: {//all rows prior to last row if(m_skip_non_aggregate_op == false || is_aggregate() == true) { - (*m_func_impl)(&arguments, &m_result); + try { + (*m_func_impl)(&arguments, &m_result); + } + catch(std::exception& e) + { + std::string error_msg = "[" + m_func_impl->m_function_name + " failed : " + std::string(e.what()) + "]"; + throw base_s3select_exception(error_msg.data(), base_s3select_exception::s3select_exp_en_t::FATAL); + } } else if(m_skip_non_aggregate_op == true) { for(auto& p : arguments) {//evaluating the arguments (not the function itself, which is a non-aggregate function) //i.e. in the following use case substring( , sum(),count() ) ; only sum() and count() are evaluated. - p->eval(); + try { + p->eval(); + } + catch(std::exception& e) + { + std::string error_msg = m_func_impl->m_function_name + " failed : " + std::string(e.what()); + throw base_s3select_exception(error_msg.data(), base_s3select_exception::s3select_exp_en_t::FATAL); + } } } } @@ -452,9 +467,27 @@ public: {//on the last row, the aggregate function is finalized, //and non-aggregate function is evaluated with the result of aggregate function. if(is_aggregate()) - (*m_func_impl).get_aggregate_result(&m_result); + { + try{ + (*m_func_impl).get_aggregate_result(&m_result); + } + catch(std::exception& e) + { + std::string error_msg = m_func_impl->m_function_name + " failed : " + std::string(e.what()); + throw base_s3select_exception(error_msg.data(), base_s3select_exception::s3select_exp_en_t::FATAL); + } + } else - (*m_func_impl)(&arguments, &m_result); + { + try{ + (*m_func_impl)(&arguments, &m_result); + } + catch(std::exception& e) + { + std::string error_msg = m_func_impl->m_function_name + " failed : " + std::string(e.what()); + throw base_s3select_exception(error_msg.data(), base_s3select_exception::s3select_exp_en_t::FATAL); + } + } } return m_result.get_value(); @@ -736,6 +769,10 @@ struct _fn_to_int : public base_function var_result = static_cast(v.dbl()); break; + case value::value_En_t::S3NULL: + var_result.setnull(); + break; + default: var_result = v.i64(); break; @@ -782,6 +819,10 @@ struct _fn_to_float : public base_function var_result = v.dbl(); break; + case value::value_En_t::S3NULL: + var_result.setnull(); + break; + default: var_result = v.i64(); break; @@ -2017,6 +2058,11 @@ struct _fn_to_bool : public base_function { i = func_arg.i64(); } + else if (func_arg.type == value::value_En_t::S3NULL) + { + result->set_null(); + return true; + } else { i = 0; @@ -2035,12 +2081,14 @@ struct _fn_to_bool : public base_function struct _fn_trim : public base_function { + //TODO base function trim std::string input_string; value v_remove; value v_input; _fn_trim() { + //default character to remove is blank v_remove = " "; } @@ -2053,13 +2101,16 @@ struct _fn_trim : public base_function { base_statement* str = *iter; v_input = str->eval(); if(v_input.type != value::value_En_t::STRING) { - throw base_s3select_exception("content is not string"); + throw base_s3select_exception("content type is not a string"); } input_string = v_input.str(); if (args_size == 2) { iter++; base_statement* next = *iter; v_remove = next->eval(); + if(v_remove.type != value::value_En_t::STRING) { + throw base_s3select_exception("remove type is not a string"); + } } boost::trim_right_if(input_string,boost::is_any_of(v_remove.str())); boost::trim_left_if(input_string,boost::is_any_of(v_remove.str())); @@ -2069,13 +2120,13 @@ struct _fn_trim : public base_function { }; struct _fn_leading : public base_function { - std::string input_string; value v_remove; value v_input; _fn_leading() { + //default character to remove is blank v_remove = " "; } @@ -2088,13 +2139,16 @@ struct _fn_leading : public base_function { base_statement* str = *iter; v_input = str->eval(); if(v_input.type != value::value_En_t::STRING) { - throw base_s3select_exception("content is not string"); + throw base_s3select_exception("content type is not a string"); } input_string = v_input.str(); if (args_size == 2) { iter++; base_statement* next = *iter; v_remove = next->eval(); + if(v_remove.type != value::value_En_t::STRING) { + throw base_s3select_exception("remove type is not a string"); + } } boost::trim_left_if(input_string,boost::is_any_of(v_remove.str())); result->set_value(input_string.c_str()); @@ -2110,6 +2164,7 @@ struct _fn_trailing : public base_function { _fn_trailing() { + //default character to remove is blank v_remove = " "; } @@ -2122,13 +2177,16 @@ struct _fn_trailing : public base_function { base_statement* str = *iter; v_input = str->eval(); if(v_input.type != value::value_En_t::STRING) { - throw base_s3select_exception("content is not string"); + throw base_s3select_exception("content type is not a string"); } input_string = v_input.str(); if (args_size == 2) { iter++; base_statement* next = *iter; v_remove = next->eval(); + if(v_remove.type != value::value_En_t::STRING) { + throw base_s3select_exception("remove type is not a string"); + } } boost::trim_right_if(input_string,boost::is_any_of(v_remove.str())); result->set_value(input_string.c_str()); @@ -2183,7 +2241,8 @@ struct _fn_decimal_operator : public base_function { iter++; base_statement* expr_scale = *iter; value expr_scale_val = expr_scale->eval(); - + + //parser does the type checking precision = expr_precision_val.i64(); scale = expr_scale_val.i64(); @@ -2195,20 +2254,20 @@ struct _fn_decimal_operator : public base_function { struct _fn_engine_version : public base_function { - const char* version_description =R"(PR #137 : -the change handle the use cases where the JSON input starts with an anonymous array/object this may cause wrong search result per the user request(SQL statement) - -handle the use-case where the user requests a json-key-path that may point to a non-discrete value. i.e. array or an object. -editorial changes. - -fix for CSV flow, in the case of a "broken row" (upon processing stream of data) - -null results upon aggregation functions on an empty group (no match for where clause). + const char* version_description =R"( +-- trim operator: case insensitive #140 +-- add exception handling to avoid crashes, and produce informative messages instead #141 +-- case-insensitive in the case of is null or is not null predicates. #141 +-- a fix for missing check-type, which cause a crash(trim operator) #142 +-- cast null operations returned false instead of null. #143 +-- adding another way to generate TPCDS data, this method is faster and efficient, it launches multiple instances of data-generators and uses less disk space #145 +-- the scripts use the dsdgen application resides on https://github.com/galsalomon66/tpc-ds-datagen-to-aws-s3 +the whole system resides in a container [ docker pull galsl/fedora_38:tpcds_v2 ] #146 +-- upon logical_operand(and/or) the parser-builder does not use case-insensitive compare function, resulting in wrong evaluation #147 )"; - _fn_engine_version() - { + {//it means it will return a single result line, in case of multi-rows input object aggregate = true; } @@ -2535,6 +2594,17 @@ bool base_statement::is_column_reference() const return false; } +std::string base_statement::get_key_from_projection() +{ + variable* v_name = dynamic_cast(this); + + if(v_name) { + return v_name->get_name(); + } else { + throw base_s3select_exception("key not present"); + } +} + bool base_statement::is_nested_aggregate(bool &aggr_flow) const { if (is_aggregate()) diff --git a/src/s3select/include/s3select_json_parser.h b/src/s3select/include/s3select_json_parser.h index aa06163f5..4b10bd732 100644 --- a/src/s3select/include/s3select_json_parser.h +++ b/src/s3select/include/s3select_json_parser.h @@ -826,4 +826,3 @@ class JsonParserHandler : public rapidjson::BaseReaderHandler, #endif - diff --git a/src/s3select/include/s3select_oper.h b/src/s3select/include/s3select_oper.h index 89544fc1d..67bfec2a7 100644 --- a/src/s3select/include/s3select_oper.h +++ b/src/s3select/include/s3select_oper.h @@ -389,7 +389,20 @@ struct binop_modulo { throw base_s3select_exception("Mod zero is not allowed"); } else { - return a % b; + return a % b; + } + } +}; + +struct binop_float_modulo +{ + double operator()(double a, double b) + { + if (b == 0) + { + throw base_s3select_exception("Mod zero is not allowed"); + } else { + return fmod(a, b); } } }; @@ -1098,8 +1111,10 @@ public: value & operator%(const value &v) { - if(v.type == value_En_t::DECIMAL) { + if(v.type == value_En_t::DECIMAL && this->type == value_En_t::DECIMAL) { return compute(*this,v); + } else if(v.type == value_En_t::FLOAT || this->type == value_En_t::FLOAT) { + return compute(*this,v); } else { throw base_s3select_exception("wrong use of modulo operation!"); } @@ -1478,6 +1493,7 @@ public: const base_statement* get_aggregate() const; bool is_nested_aggregate(bool&) const; bool is_column_reference() const; + std::string get_key_from_projection(); bool mark_aggreagtion_subtree_to_execute(); bool is_statement_contain_star_operation() const; void push_for_cleanup(std::set&); @@ -1714,7 +1730,7 @@ public: virtual bool is_column() const //is reference to column. { - if(m_var_type == var_t::VARIABLE_NAME || m_var_type == var_t::POS || m_var_type == var_t::STAR_OPERATION) + if(m_var_type == var_t::VARIABLE_NAME || m_var_type == var_t::POS || m_var_type == var_t::STAR_OPERATION || m_var_type == var_t::JSON_VARIABLE) { return true; } @@ -1817,7 +1833,7 @@ public: } } - } + } if (m_projection_alias) { @@ -2375,6 +2391,17 @@ class base_date_diff : public base_function ptime1 += boost::posix_time::minutes(ts1_td.minutes() * -1); ptime2 = ts2_ptime + boost::posix_time::hours(ts2_td.hours() * -1); ptime2 += boost::posix_time::minutes(ts2_td.minutes() * -1); + + try{ + ptime1.date().year(); + ptime2.date().year(); + } + catch(std::exception& e) + { + std::string error_msg = "the input timestamp for the diff operation is not correct : " + std::string(e.what()); + throw base_s3select_exception(error_msg.data(),base_s3select_exception::s3select_exp_en_t::FATAL); + } + } }; -- cgit v1.2.3