DOC-2200 supporting null val #308
Conversation
| .NULL Value Ingestion | ||
| + | ||
| The `INSERT INTO` statement supports inserting `NULL` values for attributes defined as `NULLABLE` in the schema. | ||
| * `NULL` or `null` can be used in the `VALUES` clause to insert a NULL value. |
There was a problem hiding this comment.
| * `NULL` or `null` can be used in the `VALUES` clause to insert a NULL value. | |
| * Either `NULL` or `null` can be used in the `VALUES` clause to represent a NULL value. |
| The optional _name_ which follows the first two (id) values is to specify the source vertex type and target vertex type, if the edge type had been defined with wildcard vertex types. | ||
| + | ||
| .NULL Value Ingestion |
There was a problem hiding this comment.
This section does not seem to be in the right place on this page.
You have it as a continuation of the section on " two options for specifying the attributes of the vertex or edge type for the values provided:" But NULLABLE is about the command in general, not about these two options.
I have more to say on this, which I will write in a separate note.
| ---- | ||
| ADD VERTEX Vertex_Type_Name ( PRIMARY_ID id_name id_type | ||
| [, attribute_name type [DEFAULT default_value] ]* ) | ||
| [, attribute_name type [NULLABLE][DEFAULT default_value] ]* ) |
There was a problem hiding this comment.
This syntax detail is correct, but there is another section on the page that needs to same treatment.
Besides the ADD VERTEX/EDGE section, there is an analogous
ADD VERTEX/EDGE (global) section.
https://docs.tigergraph.com/gsql-ref/4.2/ddl-and-loading/modifying-a-graph-schema#_add_vertex_edge_global
| ] | ||
| ---- | ||
|
|
||
| ==== Loading NULL during creation of loading job |
There was a problem hiding this comment.
I would modify and reduce the scope of this section.
There are basically 3 ways to end up with NULL in the loaded data:
- Because NULL (including empty fields) was in the input data.
That is already covered in the section above.
- Because the loading job has an explicit value (NULL) instead of making a reference (e.g. $1) to a source file.
This can happen with any data type (explicit INT, explicit BOOLEAN, ...), so there does not seem to be a reason to make an exception for NULL.
- Because the loading job uses an underscore "_" to skip an attribute.
This is discussed but not that clearly, and NULL changes the odd behavior, so we should write about this.
We might want a section called "Omitted Fields and Default Values".
Here we would:
- Talk about omitted fields in CSV, JSON, and loading job "destination clauses": if they are legal and how they are interpreted. If legal, then the default value is used.
- Explain the rules for default values.
** Depends if the particular vertex or edge type defined a custom default value for that attribute
** Depends on whether the attribute is NULLABLE or not
| [NOTE] | ||
| ==== | ||
| Primitive data types used as attributes of a vertex or edge, can be set to `NULL`. | ||
| ==== |
There was a problem hiding this comment.
I would make this a new (brief) section on Null Values, not a Note.
- If an attribute is defined as NULLABLE, then its value can be NULL, in addition to its characteristic values
- explain what it means if an attribute is null
- explain how to format null in as an input
| ** Keys and Values cannot be `NULL`. | ||
| ==== | ||
|
|
There was a problem hiding this comment.
Not all of this belongs in a special note:
LIST,SET,MAPcan be assignedNULLas an attribute
This is just following the standard rule that any attribute that is NULLABLE can have null. No special treatment.
** Elements inside
LISTorSETcannot beNULL.
This does deserve being in a NOTE because this is where TigerGraph's support is not as complete as it is in some other products.
- For
MAP:
** Keys and Values cannot beNULL
I would just lump this together with LIST and SET.
**
NULLis different from [NULL, NULL] or {NULL}.
From a technical standpoint, this is not saying anything that isn't already said in the preceding rules. It's usefulness is as an example, not an additional rule.
MyList = NULL // This is setting the entire list to null. Supported.
MyList = [NULL] // This is creating a list with one element whose value is null. Not supported.
| <1> File path specified at compile time. | ||
| <2> Run-time specification will override path specified at compile time. | ||
|
|
||
| === Loading NULL Values |
There was a problem hiding this comment.
We already have sections that talk about how to load each type of data. It is called "Attributes and attribute expressions", with a subsection for each data type.
I would move this section there, and follow the format used there.
No description provided.