-
Notifications
You must be signed in to change notification settings - Fork 3
Description
Describe the bug
This issue should possibly be in the intake-dataframe-catalog repo, but I'm opening here because the demo of the behaviour uses the ACCESS-NRI Intake Catalog.
Basically, releases of the conda/analysis3 environment more recent that 25.07 produce unexpected behaviour when using the require_all=True argument on .search.
To Reproduce
Run the following:
import intake
catalog = intake.cat.access_nri
variable=[
"tx_trans", "ty_trans",
"mld",
"area_t",
"^d[hz]t$",
]
catalog.search(
model="ACCESS-OM2.*",
variable=variable,
require_all=True
)With conda/analysis3-25.07 this produces a filtered catalog containing only the (46) experiments that include all the required variables
With conda/analysis3>25.07 the search appears to treat every variable as a regex and so includes experiments that don't contain all the required variables. It doesn't appear to be that simple though, as the following returns no experiments:
variable=[
"^tx_trans$", "^ty_trans$",
"^mld$",
"^area_t$",
"^d[hz]t$",
]
catalog.search(
model="ACCESS-OM2.*",
variable=variable,
require_all=True
)Additional context
I came across this behaviour while responding to this Forum topic.
Metadata
Metadata
Assignees
Labels
Type
Projects
Status