CSV data file format for importing data into a Graph Database (GDB) instance-Graph Database(Graph Database)-阿里云帮助中心

This topic describes the Comma-Separated Values (CSV) data file format for importing data into a Graph Database (GDB) instance.

Background information

CSV data files follow the RFC 4180 specification. For more information, see Common Format and MIME Type for Comma-Separated Values (CSV) Files.

Important

All files must be encoded in UTF-8 format.

Usage notes

Place vertices and edges in separate files.
Note
The loader can load multiple vertex files and multiple edge files in a single load task.
For each load task, all files to be loaded must be in the same folder of a bucket. You cannot use file names or file name extensions to distinguish vertex files from edge files.
Each file must contain a header row. This row consists of comma-separated system and property column headings.
The header row is the first row of the data file and ends with \n or \r\n. The row does not support special characters such as double quotation marks and line feeds. All characters except the comma (,) separator are parsed as header content.

System column headings

Header type	Parameter/Syntax	Required	Description
Point Title	`~id`	Yes	The ID of the vertex.
Point Title	`~label`	Yes	The label of the vertex. Only a single label is supported.
Edge header	`~id`	Yes	The ID of the edge.
	`~from`	Yes	The ID of the source vertex.
	`~to`	Yes	The ID of the destination vertex.
	`~label`	Yes	The label of the edge. Only a single label is supported.

Property column headings

To specify an attribute column for a regular property, use the following syntax. The type name is case-insensitive.
```
propertyname:type
```
Important
- Column headings cannot contain spaces. Therefore, property names cannot contain spaces.
- Properties do not support the List data type. A property can have only one value.
Example: Specify the column heading for a property named age of the Int type.
Each row in the file must have an integer value in this column or be left empty.
```
age:Int
```

To specify an attribute column for a SET property, use one of the following syntaxes.

Important

In Graph Database (GDB), only vertices support SET properties. Edges do not support SET properties. An error occurs if you set a SET property for an edge.

# Compatible with the default format for regular single-value properties.
propertyname:type

# Regular single-value property.
propertyname:type(single)

# SET property. The data row contains a single value.
propertyname:type(set)

# SET property. The data row contains multiple values.
propertyname:type[]

# SET property. The data row contains multiple values.
propertyname:type(set)[]

Row format

Item	Description
Separators	Fields in a row are separated by commas. Rows are separated by line feeds (`\n`) or carriage return line feeds (`\r\n`).
Blank fields	Blank fields are allowed for non-required columns, such as user-defined properties. Blank fields still require comma separators.
Vertex ID	The `~id` values of all vertices in each vertex file must be unique. Multiple vertex rows that have the same `~id` value apply to a single vertex in the graph.
Edge ID	The `~id` values of all edges in each edge file must be unique. Multiple edge rows that have the same `~id` value apply to a single edge in the graph.
Labels	Labels are case-sensitive strings.
String values	Quotation marks are optional. If a string contains special characters such as commas, you must enclose the string in double quotation marks ("").
`propertyname:type(set)`	SET property field. This indicates that the data field is a property value of the `propertyname` property and the type is `type`. Important When you overwrite a SET property, a property value is appended. If the original property is a regular property, it becomes a SET property after the update.
`propertyname:type(set)[]`	SET property field. This indicates that the data field is an array of property values for the `propertyname` property. The array elements are of the `type` type. Array elements are separated by semicolons (;).

Property overwrite logic

Original property logic	New property logic	Result
Regular property	Regular property	The original property value is overwritten. The last written value is retained.
Regular property	SET property	The property is updated to a SET property. The original property value is merged as a value in the SET.
SET property	SET property	The SET property is appended. The property values are the union of the old and new values.
SET property	Regular property	The property is updated to a regular property. The property value is the last written value.

CSV file format specifications

The format specifications for SET properties are as follows:

Single-value field for a SET property: The format is the same as the original parsing format. The parsed value is treated as a single element of the SET property.
Array field for a SET property: Use semicolons (;) to separate array elements. For example, an integer array field can be 32;45;12.
Array field for a SET property: You can enclose the entire field in double quotation marks (""). The double quotation marks at the beginning and end of the field are removed during parsing. For example, an integer array field can be "32;45;12".
If an array field for a SET property is a string, an array element cannot contain a semicolon (;) because semicolons are parsed as element separators.

Data types

Property value type	Field type	Value range
Bool or Boolean	Boolean	FALSE or false TRUE or true
Integer	char	-128 to 127
	short	-32768 to 32767
	int	-2^31 to 2^31 - 1
	long	-2^63 to 2^63 - 1
Decimal Note Decimal notation or scientific notation is supported. (+/-)INFINITY or NaN is not supported.	float	32-bit IEEE 754 floating-point
	double	64-bit IEEE 754 floating-point
String	string	Double quotation marks are optional. If a string contains special characters, such as commas (,), line feeds (`\r\n` or `\n`), or double quotation marks (""), you must enclose the string in double quotation marks (""). Example: "Hello,World" If a string enclosed in double quotation marks contains double quotation marks, escape the inner double quotation marks with two double quotation marks based on CSV rules. Example: "Hello""World""" If a string does not contain special characters, you do not need to enclose it in double quotation marks. Zero-length empty strings are supported and are represented as "". A blank field that is not enclosed in double quotation marks indicates that no value exists and is treated as `nullptr`.

General examples

Graph Database (GDB) data file (TinkerPop modern graph)
The CSV data files are as follows:
Dotfile
~id
~label
name:string
lang:string
age:int
1
person
marko
29
2
person
vadas
27
3
software
lop
java
4
person
josh
32
5
software
ripple
java
6
person
peter
35
Edge file
~id
~from
~to
~label
weight:double
7
1
2
knows
0.5
8
1
4
knows
1.0
9
1
3
created
0.4
10
4
5
created
1.0
11
4
3
created
0.4
12
6
3
created
0.2

SET property examples

CSV data file

~id	~label	fruits:string(set)[]	codes:long(set)[]
1	person	apple;pear	22
2	person	banana;bitterorange	25;81
3	person	"cherry;blackberry;grape"	3;12
4	person		"54;11"

~id	~label	name:string	lang:string	age:int
1	person	marko		29
2	person	vadas		27
3	software	lop	java
4	person	josh		32
5	software	ripple	java
6	person	peter		35

~id	~from	~to	~label	weight:double
7	1	2	knows	0.5
8	1	4	knows	1.0
9	1	3	created	0.4
10	4	5	created	1.0
11	4	3	created	0.4
12	6	3	created	0.2