https://pulumi.com logo
Title
g

gentle-account-13294

02/26/2022, 6:45 PM
pulumi.Outputs
question tied to
resource_name
: I would like to have a
resource_name
by dynamic and tied to a
function
that is taking a
Output
as parameter. e.g.
def some_func(database_name: Output[str], schema: Output[str]) :
   snowflake.SchemaGrant(f"database_{database_name}_schema_{schema}_privilege_{one_privilege}_future_false_{role_hash}",
                database_name=database_name,
                privilege=one_privilege,
                roles=roles,
                schema_name=schema,
                with_grant_option=False)
In this specific case I’m trying to write a generic function that will creates schema’s using the Snowflake provider but in principle this question should be applicable to any resource. It seems like the
resource_name
should be a
str
and using the
Output
just causes the resource name to have something like:
<pulumi.output.Output object at 0x111f482d0>_schema_<pulumi.output.Output object at 0x111f51290>_privilege_CREATE
instead of the actual name of the Output. Note that the Output is already known in this case. I also tried using
apply
For e.g.
all_outputs = pulumi.Output.all(database_name, schema)
    all_outputs.apply(inner). # where inner function just encapsulates the real work
the only way i could think of is using two
apply
one inside the other for the two outputs to resolve this. I’m hoping there is a better way or I’m missing something in my understanding of using
Outputs
??
i guess another way to ask my question would be, can resource names be dynamic and dependent on Output of previous resources ?
Output.all
seems to have worked, though would be good to discuss with experts here whether this is the correct approach.. e.g.
def snowflake_db_table_grant_future_access(database_name: Output[str], schema: Output[str], privileges: list, roles: list):
    """Grant access to a list of privileges for a schema to a list of roles ....
    """

    def inner(all_outputs: Output[List[Output]]):

        
        inner_database_name = all_outputs[0]
        inner_schema = all_outputs[1]

        role_hash = abs(hash(str(roles)))
        for one_privilege in privileges:
            snowflake.TableGrant(f"{inner_database_name}_schema_{inner_schema}_table_privilege_{one_privilege}_future_true_{role_hash}",
                database_name=database_name,
                privilege=one_privilege,
                roles=roles,
                schema_name=schema,
                on_future=True
                )
    
    all_outputs = pulumi.Output.all(database_name, schema)
    all_outputs.apply(inner)
when i call the above function , resources are now created with correct values for the
database_name
and
schema_name
.. though I guess in really they can even be thought of as CONSTANTS 🤷
e

echoing-dinner-19531

02/26/2022, 8:13 PM
i guess another way to ask my question would be, can resource names be dynamic and dependent on Output of previous resources ?
Not really. You can put the whole resource inside an apply but that has problems with previews and update plans.
There's an open issue about this: https://github.com/pulumi/pulumi/issues/5234
👀 1
g

gentle-account-13294

02/26/2022, 10:42 PM
I see,. I might abondon this altogether then 😟