This archive contains answers to questions sent to Unidata support through mid-2025. Note that the archive is no longer being updated. We provide the archive for reference; many of the answers presented here remain technically correct, even if somewhat outdated. For the most up-to-date information on the use of NSF Unidata software and data services, please consult the Software Documentation first.
Peter, The time of a data-product that is computed using a day-of-the-month field is constrained by the heuristic to be less than the creation-time of the product plus 1.5 days. The 1.5 day threshold was based on our experience with mislabeled WMO products. The labeling of your products appears to be intentional rather than accidental, so it's not surprising that the heuristic fails. I could add an option to pqact(1) that would modify the threshold. Is there no way, however, to get the data-products correctly labeled? Their identifiers aren't WMO headers -- so they're not following a strong standard. > I couldn't find anything in the support archives on this. Sorry if it's a > deja vu question... > > I'm storing model data ingested with ldm in files that include the valid > date. The product header includes the valid month and day, but not the > year. I presumed that by using the (\n:yyyy) where the \n refers to the > parenthesized valid day, that yyyy would be set to the next year if the day > number was a low value and the current month is December. > > For example, if on Dec 28th 2014, I get a model file that is valid on Jan > 2nd, then pqact would set the (\1:yyyy) field to 2015 if \n referred to the > field containing the value "02". But in my setup, I'm getting 2014 instead > ?. > > ? > Here's a > ?n (pared down) ? > sample from my pqact capturing some ECMWF from a internal > ?server? > . > ? (I know my paste converted TABs to whitespace in this example below).? > > WSI > ? ^ecmwf.(S1D|HFE)([0-1][0-3])([0-3]?[0-9])([0-2][0-9])00.* FILE > -close /ecmwf/(\3:yyyy)\2\3\4.\1.grib > > where a typical incoming header on Dec 28th might be ecmwf.S1D01021200.grib > indicating the valid date of this grib is Jan 2nd 12 UTC. In this case, > the grib data gets stored in a file called /ecmwf/2014010212.S1D.grib, > whereas I would have expected it to be /ecmwf/2015010212.S1D.grib. (To > make it easier to follow, I stripped out the part of the product > header/name that includes the initialization date/time in this example) > > ?On what day of the month will pqact set the yyyy field to the next year? > Is that configurable? > > Thanks for any insight. > > Regards, > > Peter Neilley > ? > > > -- > * Peter **Neilley *| > * w:* 978 983 6554 *e:* address@hidden Regards, Steve Emmerson Ticket Details =================== Ticket ID: BFP-983048 Department: Support LDM Priority: Normal Status: Closed