Parser handles single date incorrectly.
alxmrs opened this issue · comments
Alex Merose commented
I have the following config:
[parameters]
client=mars
target_path=pressure-{}T00z.gb
partition_keys=
date
[selection]
# Operational data
class=od
# Operational Archive
stream=oper
# Operational version
expver=1
# Forecasts
type=fc
# Priority range
date=2016-01-01
# Model levels
levtype=pl
# Initialization times
time=00
# Prevent caching of the data (recommended for batch access)
# https://confluence.ecmwf.int/display/UDOC/Retrieve#Retrieve-Retrievalefficiency
use=infrequent
# Grid in the archived format
# https://confluence.ecmwf.int/pages/viewpage.action?pageId=123799065
grid=av
# WeatherBench pressure levels
levelist=50/100/150/200/250/300/400/500/600/700/850/925/1000
step=0/1/2/3/4/5/6/7/8/9/10/11/12/13/14/15/16/17/18/19/20/21/22/23
# All available params
param=60.128/129.128/130.128/131/132/133.128/135.128/138.128/155.128/157.128/203.128
When I run this with a dry-run, it shows that it will try to write multiple shards (I don't totally understand the pattern). However, I expect this to produce only one shard: 2016-01-01
.
Workaround
I can use the range syntax to provide a range of a single date:
date=2016-01-01/to/2016-01-01
Alex Merose commented
Fixed in #144.