Merge remote-tracking branch 'origin/master' into ADQM-976

docs/en/interfaces/formats.md

@@ -471,6 +471,7 @@ The CSV format supports the output of totals and extremes the same way as `TabSe
 - [input_format_csv_skip_trailing_empty_lines](/docs/en/operations/settings/settings-formats.md/#input_format_csv_skip_trailing_empty_lines) - skip trailing empty lines at the end of data. Default value - `false`.
 - [input_format_csv_trim_whitespaces](/docs/en/operations/settings/settings-formats.md/#input_format_csv_trim_whitespaces) - trim spaces and tabs in non-quoted CSV strings. Default value - `true`.
 - [input_format_csv_allow_whitespace_or_tab_as_delimiter](/docs/en/operations/settings/settings-formats.md/# input_format_csv_allow_whitespace_or_tab_as_delimiter) - Allow to use whitespace or tab as field delimiter in CSV strings. Default value - `false`.
+- [input_format_csv_allow_variable_number_of_columns](/docs/en/operations/settings/settings-formats.md/#input_format_csv_allow_variable_number_of_columns) - ignore extra columns in CSV input (if file has more columns than expected) and treat missing fields in CSV input as default values. Default value - `false`.
 
 ## CSVWithNames {#csvwithnames}
 

docs/en/operations/settings/settings-formats.md

@@ -931,6 +931,11 @@ Result
 ```text
 "  string  "
 ```
+### input_format_csv_allow_variable_number_of_columns {#input_format_csv_allow_variable_number_of_columns}
+
+ignore extra columns in CSV input (if file has more columns than expected) and treat missing fields in CSV input as default values.
+
+Disabled by default.
 
 ### input_format_csv_allow_whitespace_or_tab_as_delimiter {#input_format_csv_allow_whitespace_or_tab_as_delimiter}
 

docs/ru/interfaces/formats.md

@@ -401,8 +401,8 @@ $ clickhouse-client --format_csv_delimiter="|" --query="INSERT INTO test.csv FOR
 - [output_format_csv_crlf_end_of_line](../operations/settings/settings.md#output_format_csv_crlf_end_of_line) - если установлено значение true, конец строки в формате вывода CSV будет `\r\n` вместо `\n`. Значение по умолчанию - `false`.
 - [input_format_csv_skip_first_lines](../operations/settings/settings.md#input_format_csv_skip_first_lines) - пропустить указанное количество строк в начале данных. Значение по умолчанию - `0`.
 - [input_format_csv_detect_header](../operations/settings/settings.md#input_format_csv_detect_header) - обнаружить заголовок с именами и типами в формате CSV. Значение по умолчанию - `true`.
-- [input_format_csv_trim_whitespaces](../operations/settings/settings.md#input_format_csv_trim_whitespaces) - удалить пробелы и символы табуляции из строк без кавычек.
-Значение по умолчанию - `true`.
+- [input_format_csv_trim_whitespaces](../operations/settings/settings.md#input_format_csv_trim_whitespaces) - удалить пробелы и символы табуляции из строк без кавычек. Значение по умолчанию - `true`.
+- [input_format_csv_allow_variable_number_of_columns](../operations/settings/settings.md/#input_format_csv_allow_variable_number_of_columns) - игнорировать дополнительные столбцы (если файл содержит больше столбцов чем ожидается) и рассматривать отсутствующие поля в CSV в качестве значений по умолчанию. Значение по умолчанию - `false`.
 
 ## CSVWithNames {#csvwithnames}
 

docs/ru/operations/settings/settings.md

@@ -1686,7 +1686,7 @@ SELECT * FROM table_with_enum_column_for_csv_insert;
 ## input_format_csv_detect_header {#input_format_csv_detect_header}
 
 Обнаружить заголовок с именами и типами в формате CSV.
- 
+
 Значение по умолчанию - `true`.
 
 ## input_format_csv_skip_first_lines {#input_format_csv_skip_first_lines}
@@ -1727,6 +1727,12 @@ echo '  string  ' | ./clickhouse local -q  "select * from table FORMAT CSV" --in
 "  string  "
 ```
 
+## input_format_csv_allow_variable_number_of_columns {#input_format_csv_allow_variable_number_of_columns}
+
+Игнорировать дополнительные столбцы (если файл содержит больше столбцов чем ожидается) и рассматривать отсутствующие поля в CSV в качестве значений по умолчанию.
+
+Выключено по умолчанию.
+
 ## output_format_tsv_crlf_end_of_line {#settings-output-format-tsv-crlf-end-of-line}
 
 Использовать в качестве разделителя строк для TSV формата CRLF (DOC/Windows стиль) вместо LF (Unix стиль).

src/Core/Settings.h

@@ -1011,6 +1011,7 @@ class IColumn;
     M(Bool, regexp_dict_allow_hyperscan, true, "Allow regexp_tree dictionary using Hyperscan library.", 0) \
     \
     M(Bool, dictionary_use_async_executor, false, "Execute a pipeline for reading from a dictionary with several threads. It's supported only by DIRECT dictionary with CLICKHOUSE source.", 0) \
+    M(Bool, input_format_csv_allow_variable_number_of_columns, false, "Ignore extra columns in CSV input (if file has more columns than expected) and treat missing fields in CSV input as default values", 0) \
 
 // End of FORMAT_FACTORY_SETTINGS
 // Please add settings non-related to formats into the COMMON_SETTINGS above.

src/Dictionaries/CacheDictionary.cpp

@@ -138,7 +138,7 @@ Columns CacheDictionary<dictionary_key_type>::getColumns(
     const Columns & default_values_columns) const
 {
     /**
-    * Flow of getColumsImpl
+    * Flow of getColumnsImpl
     * 1. Get fetch result from storage
     * 2. If all keys are found in storage and not expired
     *   2.1. If storage returns fetched columns in order of keys then result is returned to client.

src/Formats/FormatFactory.cpp

@@ -72,6 +72,7 @@ FormatSettings getFormatSettings(ContextPtr context, const Settings & settings)
     format_settings.csv.skip_trailing_empty_lines = settings.input_format_csv_skip_trailing_empty_lines;
     format_settings.csv.trim_whitespaces = settings.input_format_csv_trim_whitespaces;
     format_settings.csv.allow_whitespace_or_tab_as_delimiter = settings.input_format_csv_allow_whitespace_or_tab_as_delimiter;
+    format_settings.csv.allow_variable_number_of_columns = settings.input_format_csv_allow_variable_number_of_columns;
     format_settings.hive_text.fields_delimiter = settings.input_format_hive_text_fields_delimiter;
     format_settings.hive_text.collection_items_delimiter = settings.input_format_hive_text_collection_items_delimiter;
     format_settings.hive_text.map_keys_delimiter = settings.input_format_hive_text_map_keys_delimiter;

src/Formats/FormatSettings.h

@@ -140,6 +140,7 @@ struct FormatSettings
         bool skip_trailing_empty_lines = false;
         bool trim_whitespaces = true;
         bool allow_whitespace_or_tab_as_delimiter = false;
+        bool allow_variable_number_of_columns = false;
     } csv;
 
     struct HiveText

src/Processors/Formats/Impl/CSVRowInputFormat.cpp

@@ -283,6 +283,11 @@ bool CSVFormatReader::parseRowEndWithDiagnosticInfo(WriteBuffer & out)
     return true;
 }
 
+bool CSVFormatReader::allowVariableNumberOfColumns()
+{
+    return format_settings.csv.allow_variable_number_of_columns;
+}
+
 bool CSVFormatReader::readField(
     IColumn & column,
     const DataTypePtr & type,
@@ -347,6 +352,12 @@ bool CSVFormatReader::checkForSuffix()
     return false;
 }
 
+bool CSVFormatReader::checkForEndOfRow()
+{
+    skipWhitespacesAndTabs(*buf, format_settings.csv.allow_whitespace_or_tab_as_delimiter);
+    return buf->eof() || *buf->position() == '\n' || *buf->position() == '\r';
+}
+
 CSVSchemaReader::CSVSchemaReader(ReadBuffer & in_, bool with_names_, bool with_types_, const FormatSettings & format_settings_)
     : FormatWithNamesAndTypesSchemaReader(
         buf,

src/Processors/Formats/Impl/CSVRowInputFormat.h

@@ -69,6 +69,9 @@ public:
     void skipRowEndDelimiter() override;
     void skipPrefixBeforeHeader() override;
 
+    bool checkForEndOfRow() override;
+    bool allowVariableNumberOfColumns() override;
+
     std::vector<String> readNames() override { return readHeaderRow(); }
     std::vector<String> readTypes() override { return readHeaderRow(); }
     std::vector<String> readHeaderRow() { return readRowImpl<true>(); }

src/Processors/Formats/RowInputFormatWithNamesAndTypes.cpp

@@ -227,7 +227,30 @@ bool RowInputFormatWithNamesAndTypes::readRow(MutableColumns & columns, RowReadE
             format_reader->skipField(file_column);
 
         if (!is_last_file_column)
+        {
+            if (format_reader->allowVariableNumberOfColumns() && format_reader->checkForEndOfRow())
+            {
+                ++file_column;
+                while (file_column < column_mapping->column_indexes_for_input_fields.size())
+                {
+                    const auto & rem_column_index = column_mapping->column_indexes_for_input_fields[file_column];
+                    columns[*rem_column_index]->insertDefault();
+                    ++file_column;
+                }
+            }
+            else
+                format_reader->skipFieldDelimiter();
+        }
+    }
+
+    if (format_reader->allowVariableNumberOfColumns() && !format_reader->checkForEndOfRow())
+    {
+        do
+        {
             format_reader->skipFieldDelimiter();
+            format_reader->skipField(1);
+        }
+        while (!format_reader->checkForEndOfRow());
     }
 
     format_reader->skipRowEndDelimiter();

src/Processors/Formats/RowInputFormatWithNamesAndTypes.h

@@ -119,6 +119,10 @@ public:
     /// Check suffix.
     virtual bool checkForSuffix() { return in->eof(); }
 
+    virtual bool checkForEndOfRow() { throw Exception(ErrorCodes::NOT_IMPLEMENTED, "Method checkForEndOfRow is not implemented"); }
+
+    virtual bool allowVariableNumberOfColumns() { return false; }
+
     const FormatSettings & getFormatSettings() const { return format_settings; }
 
     virtual void setReadBuffer(ReadBuffer & in_) { in = &in_; }

tests/queries/0_stateless/00301_csv.reference

@@ -1,13 +1,32 @@
+=== Test input_format_csv_empty_as_default
 Hello, world	123	2016-01-01
 Hello, "world"	456	2016-01-02
 Hello "world"	789	2016-01-03
 Hello\n world	100	2016-01-04
 default	1	2019-06-19
 default-eof	1	2019-06-19
+=== Test datetime
 2016-01-01 01:02:03	1
 2016-01-02 01:02:03	2
 2017-08-15 13:15:01	3
 1970-01-02 05:46:39	4
+=== Test nullable datetime
 2016-01-01 01:02:03	NUL
 2016-01-02 01:02:03	Nhello
 \N	\N
+=== Test ignore extra columns
+Hello	1	String1
+Hello	2	String2
+Hello	3	String3
+Hello	4	String4
+Hello	5	String5
+Hello	6	String6
+=== Test missing as default
+	0	0	33	\N	55	Default
+	0	0	33	\N	55	Default
+Hello	0	0	33	\N	55	Default
+Hello	0	0	33	\N	55	Default
+Hello	1	3	2	\N	55	Default
+Hello	1	4	2	3	4	String
+Hello	1	4	2	3	4	String
+Hello	1	5	2	3	4	String

tests/queries/0_stateless/00301_csv.sh

@@ -4,6 +4,7 @@ CURDIR=$(cd "$(dirname "${BASH_SOURCE[0]}")" && pwd)
 # shellcheck source=../shell_config.sh
 . "$CURDIR"/../shell_config.sh
 
+echo === Test input_format_csv_empty_as_default
 $CLICKHOUSE_CLIENT --query="DROP TABLE IF EXISTS csv";
 $CLICKHOUSE_CLIENT --query="CREATE TABLE csv (s String, n UInt64 DEFAULT 1, d Date DEFAULT '2019-06-19') ENGINE = Memory";
 
@@ -18,6 +19,7 @@ Hello "world", 789 ,2016-01-03
 $CLICKHOUSE_CLIENT --query="SELECT * FROM csv ORDER BY d, s";
 $CLICKHOUSE_CLIENT --query="DROP TABLE csv";
 
+echo === Test datetime
 $CLICKHOUSE_CLIENT --query="CREATE TABLE csv (t DateTime('Asia/Istanbul'), s String) ENGINE = Memory";
 
 echo '"2016-01-01 01:02:03","1"
@@ -28,7 +30,7 @@ echo '"2016-01-01 01:02:03","1"
 $CLICKHOUSE_CLIENT --query="SELECT * FROM csv ORDER BY s";
 $CLICKHOUSE_CLIENT --query="DROP TABLE csv";
 
-
+echo === Test nullable datetime
 $CLICKHOUSE_CLIENT --query="CREATE TABLE csv (t Nullable(DateTime('Asia/Istanbul')), s Nullable(String)) ENGINE = Memory";
 
 echo 'NULL, NULL
@@ -37,3 +39,32 @@ echo 'NULL, NULL
 
 $CLICKHOUSE_CLIENT --query="SELECT * FROM csv ORDER BY s NULLS LAST";
 $CLICKHOUSE_CLIENT --query="DROP TABLE csv";
+
+
+echo === Test ignore extra columns
+$CLICKHOUSE_CLIENT --query="CREATE TABLE csv (s String, n UInt64 DEFAULT 3, d String DEFAULT 'String4') ENGINE = Memory";
+
+echo '"Hello", 1, "String1" 
+"Hello", 2, "String2",
+"Hello", 3, "String3", "2016-01-13"
+"Hello", 4,        , "2016-01-14"
+"Hello", 5, "String5", "2016-01-15", "2016-01-16"
+"Hello", 6, "String6" , "line with a
+break"' | $CLICKHOUSE_CLIENT --input_format_defaults_for_omitted_fields=1 --input_format_csv_empty_as_default=1 --input_format_csv_allow_variable_number_of_columns=1 --query="INSERT INTO csv FORMAT CSV";
+$CLICKHOUSE_CLIENT --query="SELECT * FROM csv ORDER BY s, n";
+$CLICKHOUSE_CLIENT --query="DROP TABLE csv";
+
+
+echo === Test missing as default
+$CLICKHOUSE_CLIENT --query="CREATE TABLE csv (f1 String, f2 UInt64, f3 UInt256, f4 UInt64 Default 33, f5 Nullable(UInt64), f6 Nullable(UInt64) Default 55, f7 String DEFAULT 'Default') ENGINE = Memory";
+
+echo '
+,
+"Hello"
+"Hello",
+"Hello", 1, 3, 2
+"Hello",1,4,2,3,4,"String"
+"Hello", 1, 4, 2, 3, 4, "String"
+"Hello", 1, 5, 2, 3, 4, "String",'| $CLICKHOUSE_CLIENT --input_format_defaults_for_omitted_fields=1 --input_format_csv_allow_variable_number_of_columns=1 --query="INSERT INTO csv FORMAT CSV";
+$CLICKHOUSE_CLIENT --query="SELECT * FROM csv ORDER BY f1, f2, f3, f4, f5 NULLS FIRST, f6, f7";
+$CLICKHOUSE_CLIENT --query="DROP TABLE csv";