Add TabSeparatedRawRowInputFormat

src/Processors/Formats/Impl/TabSeparatedRawRowInputFormat.h

@@ -0,0 +1,58 @@
+#pragma once
+
+#include <Core/Block.h>
+#include <Formats/FormatSettings.h>
+#include <IO/ReadBufferFromString.h>
+#include <Processors/Formats/RowInputFormatWithDiagnosticInfo.h>
+
+
+namespace DB
+{
+
+/** A stream to input data in tsv format, but without escaping individual values.
+  * It only supports one string column
+  */
+class TabSeparatedRawRowInputFormat : public TabSeparatedRowInputFormat
+{
+public:
+    /** with_names - the first line is the header with the names of the columns
+      * with_types - on the next line header with type names
+      */
+    TabSeparatedRawRowInputFormat(
+        const Block & header_,
+        ReadBuffer & in_,
+        const Params & params_,
+        bool with_names_,
+        bool with_types_,
+        const FormatSettings & format_settings_)
+        : TabSeparatedRowInputFormat(header_, in_, params_, with_names_, with_types_, format_settings_)
+    {
+    }
+
+    String getName() const override { return "TabSeparatedRawRowInputFormat"; }
+
+    bool readField(IColumn & column, const DataTypePtr & type, bool) override
+    {
+        // TODO: possible to optimize
+        std::string buf;
+
+        while (!in.eof())
+        {
+            char c = *in.position();
+
+            if (c == '\n' || c == '\t')
+                break;
+
+            in.ignore();
+            buf.push_back(c);
+        }
+
+        ReadBufferFromString line_in(buf);
+
+        type->deserializeAsWholeText(column, line_in, format_settings);
+
+        return true;
+    }
+};
+
+}

src/Processors/Formats/Impl/TabSeparatedRowInputFormat.cpp

@@ -3,6 +3,7 @@
 #include <IO/Operators.h>
 
 #include <Processors/Formats/Impl/TabSeparatedRowInputFormat.h>
+#include <Processors/Formats/Impl/TabSeparatedRawRowInputFormat.h>
 #include <Formats/verbosePrintString.h>
 #include <Formats/FormatFactory.h>
 #include <DataTypes/DataTypeNothing.h>
@@ -360,6 +361,18 @@ void registerInputFormatProcessorTabSeparated(FormatFactory & factory)
         });
     }
 
+    for (const auto * name : {"TabSeparatedRaw", "TSVRaw"})
+    {
+        factory.registerInputFormatProcessor(name, [](
+            ReadBuffer & buf,
+            const Block & sample,
+            IRowInputFormat::Params params,
+            const FormatSettings & settings)
+        {
+            return std::make_shared<TabSeparatedRawRowInputFormat>(sample, buf, params, false, false, settings);
+        });
+    }
+
     for (const auto * name : {"TabSeparatedWithNames", "TSVWithNames"})
     {
         factory.registerInputFormatProcessor(name, [](

src/Processors/Formats/Impl/TabSeparatedRowInputFormat.h

@@ -28,10 +28,14 @@ public:
 
     void resetParser() override;
 
-private:
+protected:
     bool with_names;
     bool with_types;
     const FormatSettings format_settings;
+
+    virtual bool readField(IColumn & column, const DataTypePtr & type, bool is_last_file_column);
+
+private:
     DataTypes data_types;
 
     using IndexesMap = std::unordered_map<String, size_t>;
@@ -43,8 +47,6 @@ private:
     std::vector<UInt8> read_columns;
     std::vector<size_t> columns_to_fill_with_default_values;
 
-    bool readField(IColumn & column, const DataTypePtr & type, bool is_last_file_column);
-
     void addInputColumn(const String & column_name);
     void setupAllColumnsByTableSchema();
     void fillUnreadColumnsWithDefaults(MutableColumns & columns, RowReadExtension & row_read_extension);