https://prefect.io
Join Slack
Is there a nature way to work with Parameters? e.g the example below will throw an exception because...
# ask-community
e
Is there a nature way to work with Parameters? e.g the example below will throw an exception because for_date has type Parameter
Copy code
with Flow("Demo") as flow:
    for_date = Parameter('for_date', required=True, default='20200101')
    for_date_str = dt.datetime.strptime(for_date, "%Y%m%d")
I think what I’d like to ask is how to extract the value of a parameter in order to use it somewhere
k
Hi @Enda Peng! I think the issue here is that 
for_date
is a deferred execution that is population at runtime while 
for_date_str
happens at run time. You would need to create a helper task to perform that operation and then return it so that the execution gets deferred as well
Copy code
@task
def string_to_time(for_date):
    return dt.datetime.strptime(for_date, "%Y%m%d")

with Flow("Demo") as flow:
    for_date = Parameter('for_date', required=True, default='20200101')
    for_date_str = string_to_time(for_date)
e
ha, that’s what I am doing, sound to be the 
nature
way
k
That is indeed the 
nature
way
e
thank you!
👍 1
j
@Kevin Kho Most times, in operation I'm querying for a date range of 24-48 hours. However, when I'm launching a new flow, I tend to back date or play around with the date ranges to make sure the flow is robust. Is there a way to use Parameters to manually set the parameters while the default is dynamically set?
When I tried to set 
Parameter('end_date', default=date().today().strptime(...))
the default date is static to when I deployed the flow.
k
Hi @Jimmy Le, maybe you can try using the context so it gets populated at runtime? prefect.context.today, and then maybe pass this to a task to format it the way you need?
Otherwise you may need to do this:
Copy code
import pendulum
date = Parameter(name="date", default="yesterday")
@task
def generate_date(date_parameter):
    if date_parameter == 'yesterday'
        return pendulum.now().subtract(days=1)
    return pendulum.parse(date_parameter)
j
Oooh nice, I like that. Never played around with any of these utilities actually. Going to dive in thanks!
Open in Slack
PreviousNext
Powered by

Prefect Community

Bring your towel and join one of the fastest growing data communities. Welcome to our second-generation open source orchestration platform, a completely rethought approach to dataflow automation.

Powered by