ClickHouse/docs/en/interfaces/formats.md
2019-02-28 10:45:30 +00:00

29 KiB
Raw Blame History

Formats for input and output data

ClickHouse can accept (INSERT) and return (SELECT) data in various formats.

The table below lists supported formats and how they can be used in INSERT and SELECT queries.

Format INSERT SELECT
TabSeparated
TabSeparatedRaw
TabSeparatedWithNames
TabSeparatedWithNamesAndTypes
CSV
CSVWithNames
Values
Vertical
JSON
JSONCompact
JSONEachRow
TSKV
Pretty
PrettyCompact
PrettyCompactMonoBlock
PrettyNoEscapes
PrettySpace
Protobuf
RowBinary
Native
Null
XML
CapnProto

TabSeparated

In TabSeparated format, data is written by row. Each row contains values separated by tabs. Each value is follow by a tab, except the last value in the row, which is followed by a line feed. Strictly Unix line feeds are assumed everywhere. The last row also must contain a line feed at the end. Values are written in text format, without enclosing quotation marks, and with special characters escaped.

This format is also available under the name TSV.

The TabSeparated format is convenient for processing data using custom programs and scripts. It is used by default in the HTTP interface, and in the command-line client's batch mode. This format also allows transferring data between different DBMSs. For example, you can get a dump from MySQL and upload it to ClickHouse, or vice versa.

The TabSeparated format supports outputting total values (when using WITH TOTALS) and extreme values (when 'extremes' is set to 1). In these cases, the total values and extremes are output after the main data. The main result, total values, and extremes are separated from each other by an empty line. Example:

SELECT EventDate, count() AS c FROM test.hits GROUP BY EventDate WITH TOTALS ORDER BY EventDate FORMAT TabSeparated``
2014-03-17      1406958
2014-03-18      1383658
2014-03-19      1405797
2014-03-20      1353623
2014-03-21      1245779
2014-03-22      1031592
2014-03-23      1046491

0000-00-00      8873898

2014-03-17      1031592
2014-03-23      1406958

Data formatting

Integer numbers are written in decimal form. Numbers can contain an extra "+" character at the beginning (ignored when parsing, and not recorded when formatting). Non-negative numbers can't contain the negative sign. When reading, it is allowed to parse an empty string as a zero, or (for signed types) a string consisting of just a minus sign as a zero. Numbers that do not fit into the corresponding data type may be parsed as a different number, without an error message.

Floating-point numbers are written in decimal form. The dot is used as the decimal separator. Exponential entries are supported, as are 'inf', '+inf', '-inf', and 'nan'. An entry of floating-point numbers may begin or end with a decimal point. During formatting, accuracy may be lost on floating-point numbers. During parsing, it is not strictly required to read the nearest machine-representable number.

Dates are written in YYYY-MM-DD format and parsed in the same format, but with any characters as separators. Dates with times are written in the format YYYY-MM-DD hh:mm:ss and parsed in the same format, but with any characters as separators. This all occurs in the system time zone at the time the client or server starts (depending on which one formats data). For dates with times, daylight saving time is not specified. So if a dump has times during daylight saving time, the dump does not unequivocally match the data, and parsing will select one of the two times. During a read operation, incorrect dates and dates with times can be parsed with natural overflow or as null dates and times, without an error message.

As an exception, parsing dates with times is also supported in Unix timestamp format, if it consists of exactly 10 decimal digits. The result is not time zone-dependent. The formats YYYY-MM-DD hh:mm:ss and NNNNNNNNNN are differentiated automatically.

Strings are output with backslash-escaped special characters. The following escape sequences are used for output: \b, \f, \r, \n, \t, \0, \', \\. Parsing also supports the sequences \a, \v, and \xHH (hex escape sequences) and any \c sequences, where c is any character (these sequences are converted to c). Thus, reading data supports formats where a line feed can be written as \n or \, or as a line feed. For example, the string Hello world with a line feed between the words instead of a space can be parsed in any of the following variations:

Hello\nworld

Hello\
world

The second variant is supported because MySQL uses it when writing tab-separated dumps.

The minimum set of characters that you need to escape when passing data in TabSeparated format: tab, line feed (LF) and backslash.

Only a small set of symbols are escaped. You can easily stumble onto a string value that your terminal will ruin in output.

Arrays are written as a list of comma-separated values in square brackets. Number items in the array are fomratted as normally, but dates, dates with times, and strings are written in single quotes with the same escaping rules as above.

NULL is formatted as \N.

TabSeparatedRaw

Differs from TabSeparated format in that the rows are written without escaping. This format is only appropriate for outputting a query result, but not for parsing (retrieving data to insert in a table).

This format is also available under the name TSVRaw.

TabSeparatedWithNames

Differs from the TabSeparated format in that the column names are written in the first row. During parsing, the first row is completely ignored. You can't use column names to determine their position or to check their correctness. (Support for parsing the header row may be added in the future.)

This format is also available under the name TSVWithNames.

TabSeparatedWithNamesAndTypes

Differs from the TabSeparated format in that the column names are written to the first row, while the column types are in the second row. During parsing, the first and second rows are completely ignored.

This format is also available under the name TSVWithNamesAndTypes.

TSKV

Similar to TabSeparated, but outputs a value in name=value format. Names are escaped the same way as in TabSeparated format, and the = symbol is also escaped.

SearchPhrase=   count()=8267016
SearchPhrase=bathroom interior design    count()=2166
SearchPhrase=yandex     count()=1655
SearchPhrase=2014 spring fashion    count()=1549
SearchPhrase=freeform photos       count()=1480
SearchPhrase=angelina jolie    count()=1245
SearchPhrase=omsk       count()=1112
SearchPhrase=photos of dog breeds    count()=1091
SearchPhrase=curtain designs        count()=1064
SearchPhrase=baku       count()=1000

NULL is formatted as \N.

SELECT * FROM t_null FORMAT TSKV
x=1	y=\N

When there is a large number of small columns, this format is ineffective, and there is generally no reason to use it. It is used in some departments of Yandex.

Both data output and parsing are supported in this format. For parsing, any order is supported for the values of different columns. It is acceptable for some values to be omitted they are treated as equal to their default values. In this case, zeros and blank rows are used as default values. Complex values that could be specified in the table are not supported as defaults.

Parsing allows the presence of the additional field tskv without the equal sign or a value. This field is ignored.

CSV

Comma Separated Values format (RFC).

When formatting, rows are enclosed in double quotes. A double quote inside a string is output as two double quotes in a row. There are no other rules for escaping characters. Date and date-time are enclosed in double quotes. Numbers are output without quotes. Values are separated by a delimiter character, which is , by default. The delimiter character is defined in the setting format_csv_delimiter. Rows are separated using the Unix line feed (LF). Arrays are serialized in CSV as follows: first the array is serialized to a string as in TabSeparated format, and then the resulting string is output to CSV in double quotes. Tuples in CSV format are serialized as separate columns (that is, their nesting in the tuple is lost).

clickhouse-client --format_csv_delimiter="|" --query="INSERT INTO test.csv FORMAT CSV" < data.csv

*By default, the delimiter is ,. See the format_csv_delimiter setting for more information.

When parsing, all values can be parsed either with or without quotes. Both double and single quotes are supported. Rows can also be arranged without quotes. In this case, they are parsed up to the delimiter character or line feed (CR or LF). In violation of the RFC, when parsing rows without quotes, the leading and trailing spaces and tabs are ignored. For the line feed, Unix (LF), Windows (CR LF) and Mac OS Classic (CR LF) types are all supported.

NULL is formatted as \N.

The CSV format supports the output of totals and extremes the same way as TabSeparated.

CSVWithNames

Also prints the header row, similar to TabSeparatedWithNames.

JSON

Outputs data in JSON format. Besides data tables, it also outputs column names and types, along with some additional information: the total number of output rows, and the number of rows that could have been output if there weren't a LIMIT. Example:

SELECT SearchPhrase, count() AS c FROM test.hits GROUP BY SearchPhrase WITH TOTALS ORDER BY c DESC LIMIT 5 FORMAT JSON
{
        "meta":
        [
                {
                        "name": "SearchPhrase",
                        "type": "String"
                },
                {
                        "name": "c",
                        "type": "UInt64"
                }
        ],

        "data":
        [
                {
                        "SearchPhrase": "",
                        "c": "8267016"
                },
                {
                        "SearchPhrase": "bathroom interior design",
                        "c": "2166"
                },
                {
                        "SearchPhrase": "yandex",
                        "c": "1655"
                },
                {
                        "SearchPhrase": "spring 2014 fashion",
                        "c": "1549"
                },
                {
                        "SearchPhrase": "freeform photos",
                        "c": "1480"
                }
        ],

        "totals":
        {
                "SearchPhrase": "",
                "c": "8873898"
        },

        "extremes":
        {
                "min":
                {
                        "SearchPhrase": "",
                        "c": "1480"
                },
                "max":
                {
                        "SearchPhrase": "",
                        "c": "8267016"
                }
        },

        "rows": 5,

        "rows_before_limit_at_least": 141137
}

The JSON is compatible with JavaScript. To ensure this, some characters are additionally escaped: the slash / is escaped as \/; alternative line breaks U+2028 and U+2029, which break some browsers, are escaped as \uXXXX. ASCII control characters are escaped: backspace, form feed, line feed, carriage return, and horizontal tab are replaced with \b, \f, \n, \r, \t , as well as the remaining bytes in the 00-1F range using \uXXXX sequences. Invalid UTF-8 sequences are changed to the replacement character <20> so the output text will consist of valid UTF-8 sequences. For compatibility with JavaScript, Int64 and UInt64 integers are enclosed in double quotes by default. To remove the quotes, you can set the configuration parameter output_format_json_quote_64bit_integers to 0.

rows The total number of output rows.

rows_before_limit_at_least The minimal number of rows there would have been without LIMIT. Output only if the query contains LIMIT. If the query contains GROUP BY, rows_before_limit_at_least is the exact number of rows there would have been without a LIMIT.

totals Total values (when using WITH TOTALS).

extremes Extreme values (when extremes is set to 1).

This format is only appropriate for outputting a query result, but not for parsing (retrieving data to insert in a table).

ClickHouse supports NULL, which is displayed as null in the JSON output.

See also the JSONEachRow format.

JSONCompact

Differs from JSON only in that data rows are output in arrays, not in objects.

Example:

{
        "meta":
        [
                {
                        "name": "SearchPhrase",
                        "type": "String"
                },
                {
                        "name": "c",
                        "type": "UInt64"
                }
        ],

        "data":
        [
                ["", "8267016"],
                ["bathroom interior design", "2166"],
                ["yandex", "1655"],
                ["fashion trends spring 2014", "1549"],
                ["freeform photo", "1480"]
        ],

        "totals": ["","8873898"],

        "extremes":
        {
                "min": ["","1480"],
                "max": ["","8267016"]
        },

        "rows": 5,

        "rows_before_limit_at_least": 141137
}

This format is only appropriate for outputting a query result, but not for parsing (retrieving data to insert in a table). See also the JSONEachRow format.

JSONEachRow

Outputs data as separate JSON objects for each row (newline delimited JSON).

{"SearchPhrase":"","count()":"8267016"}
{"SearchPhrase": "bathroom interior design","count()": "2166"}
{"SearchPhrase":"yandex","count()":"1655"}
{"SearchPhrase":"2014 spring fashion","count()":"1549"}
{"SearchPhrase":"freeform photo","count()":"1480"}
{"SearchPhrase":"angelina jolie","count()":"1245"}
{"SearchPhrase":"omsk","count()":"1112"}
{"SearchPhrase":"photos of dog breeds","count()":"1091"}
{"SearchPhrase":"curtain designs","count()":"1064"}
{"SearchPhrase":"baku","count()":"1000"}

Unlike the JSON format, there is no substitution of invalid UTF-8 sequences. Any set of bytes can be output in the rows. This is necessary so that data can be formatted without losing any information. Values are escaped in the same way as for JSON.

For parsing, any order is supported for the values of different columns. It is acceptable for some values to be omitted they are treated as equal to their default values. In this case, zeros and blank rows are used as default values. Complex values that could be specified in the table are not supported as defaults, but it can be turned on by option insert_sample_with_metadata=1. Whitespace between elements is ignored. If a comma is placed after the objects, it is ignored. Objects don't necessarily have to be separated by new lines.

Native

The most efficient format. Data is written and read by blocks in binary format. For each block, the number of rows, number of columns, column names and types, and parts of columns in this block are recorded one after another. In other words, this format is "columnar" it doesn't convert columns to rows. This is the format used in the native interface for interaction between servers, for using the command-line client, and for C++ clients.

You can use this format to quickly generate dumps that can only be read by the ClickHouse DBMS. It doesn't make sense to work with this format yourself.

Null

Nothing is output. However, the query is processed, and when using the command-line client, data is transmitted to the client. This is used for tests, including productivity testing. Obviously, this format is only appropriate for output, not for parsing.

Pretty

Outputs data as Unicode-art tables, also using ANSI-escape sequences for setting colors in the terminal. A full grid of the table is drawn, and each row occupies two lines in the terminal. Each result block is output as a separate table. This is necessary so that blocks can be output without buffering results (buffering would be necessary in order to pre-calculate the visible width of all the values).

NULL is output as ᴺᵁᴸᴸ.

Example (shown for the PrettyCompact format):

SELECT * FROM t_null
┌─x─┬────y─┐
│ 1 │ ᴺᵁᴸᴸ │
└───┴──────┘

Rows are not escaped in Pretty* formats. Example is shown for the PrettyCompact format:

SELECT 'String with \'quotes\' and \t character' AS Escaping_test
┌─Escaping_test────────────────────────┐
│ String with 'quotes' and 	 character │
└──────────────────────────────────────┘

To avoid dumping too much data to the terminal, only the first 10,000 rows are printed. If the number of rows is greater than or equal to 10,000, the message "Showed first 10 000" is printed. This format is only appropriate for outputting a query result, but not for parsing (retrieving data to insert in a table).

The Pretty format supports outputting total values (when using WITH TOTALS) and extremes (when 'extremes' is set to 1). In these cases, total values and extreme values are output after the main data, in separate tables. Example (shown for the PrettyCompact format):

SELECT EventDate, count() AS c FROM test.hits GROUP BY EventDate WITH TOTALS ORDER BY EventDate FORMAT PrettyCompact
┌──EventDate─┬───────c─┐
│ 2014-03-17 │ 1406958 │
│ 2014-03-18 │ 1383658 │
│ 2014-03-19 │ 1405797 │
│ 2014-03-20 │ 1353623 │
│ 2014-03-21 │ 1245779 │
│ 2014-03-22 │ 1031592 │
│ 2014-03-23 │ 1046491 │
└────────────┴─────────┘

Totals:
┌──EventDate─┬───────c─┐
│ 0000-00-00 │ 8873898 │
└────────────┴─────────┘

Extremes:
┌──EventDate─┬───────c─┐
│ 2014-03-17 │ 1031592 │
│ 2014-03-23 │ 1406958 │
└────────────┴─────────┘

PrettyCompact

Differs from Pretty in that the grid is drawn between rows and the result is more compact. This format is used by default in the command-line client in interactive mode.

PrettyCompactMonoBlock

Differs from PrettyCompact in that up to 10,000 rows are buffered, then output as a single table, not by blocks.

PrettyNoEscapes

Differs from Pretty in that ANSI-escape sequences aren't used. This is necessary for displaying this format in a browser, as well as for using the 'watch' command-line utility.

Example:

watch -n1 "clickhouse-client --query='SELECT event, value FROM system.events FORMAT PrettyCompactNoEscapes'"

You can use the HTTP interface for displaying in the browser.

PrettyCompactNoEscapes

The same as the previous setting.

PrettySpaceNoEscapes

The same as the previous setting.

PrettySpace

Differs from PrettyCompact in that whitespace (space characters) is used instead of the grid.

RowBinary

Formats and parses data by row in binary format. Rows and values are listed consecutively, without separators. This format is less efficient than the Native format, since it is row-based.

Integers use fixed-length little endian representation. For example, UInt64 uses 8 bytes. DateTime is represented as UInt32 containing the Unix timestamp as the value. Date is represented as a UInt16 object that contains the number of days since 1970-01-01 as the value. String is represented as a varint length (unsigned LEB128), followed by the bytes of the string. FixedString is represented simply as a sequence of bytes.

Array is represented as a varint length (unsigned LEB128), followed by successive elements of the array.

For NULL support, an additional byte containing 1 or 0 is added before each Nullable value. If 1, then the value is NULL and this byte is interpreted as a separate value. If 0, the value after the byte is not NULL.

RowBinaryWithNamesAndTypes

Similar to RowBinary, but with added header:

  • LEB128-encoded number of columns (N)
  • N Strings specifying column names
  • N Strings specifying column types

Values

Prints every row in brackets. Rows are separated by commas. There is no comma after the last row. The values inside the brackets are also comma-separated. Numbers are output in decimal format without quotes. Arrays are output in square brackets. Strings, dates, and dates with times are output in quotes. Escaping rules and parsing are similar to the TabSeparated format. During formatting, extra spaces aren't inserted, but during parsing, they are allowed and skipped (except for spaces inside array values, which are not allowed). NULL is represented as NULL.

The minimum set of characters that you need to escape when passing data in Values format: single quotes and backslashes.

This is the format that is used in INSERT INTO t VALUES ..., but you can also use it for formatting query results.

Vertical

Prints each value on a separate line with the column name specified. This format is convenient for printing just one or a few rows, if each row consists of a large number of columns.

NULL is output as ᴺᵁᴸᴸ.

Example:

SELECT * FROM t_null FORMAT Vertical
Row 1:
──────
x: 1
y: ᴺᵁᴸᴸ

Rows are not escaped in Vertical format:

SELECT 'string with \'quotes\' and \t with some special \n characters' AS test FORMAT Vertical
Row 1:
──────
test: string with 'quotes' and 	 with some special
 characters

This format is only appropriate for outputting a query result, but not for parsing (retrieving data to insert in a table).

XML

XML format is suitable only for output, not for parsing. Example:

<?xml version='1.0' encoding='UTF-8' ?>
<result>
        <meta>
                <columns>
                        <column>
                                <name>SearchPhrase</name>
                                <type>String</type>
                        </column>
                        <column>
                                <name>count()</name>
                                <type>UInt64</type>
                        </column>
                </columns>
        </meta>
        <data>
                <row>
                        <SearchPhrase></SearchPhrase>
                        <field>8267016</field>
                </row>
                <row>
                        <SearchPhrase>bathroom interior design</SearchPhrase>
                        <field>2166</field>
                </row>
                <row>
                        <SearchPhrase>yandex</SearchPhrase>
                        <field>1655</field>
                </row>
                <row>
                        <SearchPhrase>2014 spring fashion</SearchPhrase>
                        <field>1549</field>
                </row>
                <row>
                        <SearchPhrase>freeform photos</SearchPhrase>
                        <field>1480</field>
                </row>
                <row>
                        <SearchPhrase>angelina jolie</SearchPhrase>
                        <field>1245</field>
                </row>
                <row>
                        <SearchPhrase>omsk</SearchPhrase>
                        <field>1112</field>
                </row>
                <row>
                        <SearchPhrase>photos of dog breeds</SearchPhrase>
                        <field>1091</field>
                </row>
                <row>
                        <SearchPhrase>curtain designs</SearchPhrase>
                        <field>1064</field>
                </row>
                <row>
                        <SearchPhrase>baku</SearchPhrase>
                        <field>1000</field>
                </row>
        </data>
        <rows>10</rows>
        <rows_before_limit_at_least>141137</rows_before_limit_at_least>
</result>

If the column name does not have an acceptable format, just 'field' is used as the element name. In general, the XML structure follows the JSON structure. Just as for JSON, invalid UTF-8 sequences are changed to the replacement character <20> so the output text will consist of valid UTF-8 sequences.

In string values, the characters < and & are escaped as < and &.

Arrays are output as <array><elem>Hello</elem><elem>World</elem>...</array>,and tuples as <tuple><elem>Hello</elem><elem>World</elem>...</tuple>.

CapnProto

Cap'n Proto is a binary message format similar to Protocol Buffers and Thrift, but not like JSON or MessagePack.

Cap'n Proto messages are strictly typed and not self-describing, meaning they need an external schema description. The schema is applied on the fly and cached for each query.

cat capnproto_messages.bin | clickhouse-client --query "INSERT INTO test.hits FORMAT CapnProto SETTINGS format_schema='schema:Message'"

Where schema.capnp looks like this:

struct Message {
  SearchPhrase @0 :Text;
  c @1 :Uint64;
}

Deserialization is effective and usually doesn't increase the system load.

See also Format Schema.

Protobuf

Protobuf - is a Protocol Buffers format.

ClickHouse supports both proto2 and proto3. Repeated/optional/required fields are supported.

This format requires an external format schema. The schema is cached between queries. Usage examples:

SELECT * FROM test.table FORMAT Protobuf SETTINGS format_schema = 'schemafile:MessageType'
cat protobuf_messages.bin | clickhouse-client --query "INSERT INTO test.table FORMAT Protobuf SETTINGS format_schema='schemafile:MessageType'"

Where the file schemafile.proto looks like this:

syntax = "proto3";

message MessageType {
  string name = 1;
  string surname = 2;
  uint32 birthDate = 3;
  repeated string phoneNumbers = 4;
};

To find the correspondence between table columns and fields of Protocol Buffers' message type ClickHouse compares their names. This comparison is case-insensitive and the characters _ (underscore) and . (dot) are considered as equal. If types of a column and a field of Protocol Buffers' message are different the necessary conversion is applied.

Nested messages are supported. For example, for the field z in the following message type

message MessageType {
  message XType {
    message YType {
      int32 z;
    };
    repeated YType y;
  };
  XType x;
};

ClickHouse tries to find a column named x.y.z (or x_y_z or X.y_Z and so on). Nested messages are suitable to input or output a nested data structures.

Default values defined in a proto2 protobuf schema like this

message MessageType {
  optional int32 result_per_page = 3 [default = 10];
}

are not applied; the table defaults are used instead of them.

Format Schema

The file name containing the format schema is set by the setting format_schema. It's required to set this setting when it is used one of the formats Cap'n Proto and Protobuf. The format schema is a combination of a file name and the name of a message type in this file, delimited by colon, e.g. schemafile.proto:MessageType. If the file has the standard extension for the format (for example, .proto for Protobuf), it can be omitted and in this case the format schema looks like schemafile:MessageType.

If you input or output data via the client the file name specified in the format schema can contain an absolute path or a path relative to the current directory on the client. If you input or output data via the HTTP interface the file name specified in the format schema should be located in the directory specified in format_schema_path in the server configuration.

Original article