[Date Prev][Date Next][Thread Prev][Thread Next][Date Index][Thread Index]

[mdf] [Fwd: Question about using "DIRfile" in DSSERVE]



Hello,

I have some questions about using DIRfile keyword in DSSERVE setting.

I  have LOTS of AREA data files are named something like
?YYYYStormName.SSSS.ddhhmm.totalHour", which are located in
?YYYY/StormName? directories.

Where:
         YYYY = current year
            SSSS    = data source (AMSU, TRMM)
            Ddhhmm = date, hour, minute

 I add the DSSERVE as following:

DSSERVE ADD LOC/TR2002 AREA DIRFILE=?2002/*/*?

So the questions are:

(1).
(Please note that there are lots of AREA file in the data GROUP
?LOC/TR2002".)   If I want to IMGDISP one of the above files, said
?2002OLGA.AMSU.261208.24" (resided on ?2002/OLGA? directory), how can I
know the position number of this file in that data GROUP, so that I can
do ?IMGDISP LOC/TR2002.pos# ? for that data file.   Is there any McIDAS
function can list both of the position number in the group and the data
file name?  Or can I do the IMGDISP (or IMGCHA...) Using the actual file
name other than the position number if I use keyword "DIRfile" to set
the data set?

(2).  The actual name of the above AREA file might be as long as 50
characters, is it accepted by McIDAS?

By the way, I use McIdas 7.8.

Thanks a lot.

Jianbin

Date: Thu, 27 Jun 2002 15:09:28 -0500
From: Russ Dengel <address@hidden>
Organization: SSEC
Subject: Re: [mdf] [Fwd: Question about using "DIRfile" in DSSERVE]

Currently, the image file name(s) are not returned by the image
servers. In the example described above, the dataset naming convention
is defined using the DIRFILE= keyword. We refer to these types of
datasets as "NAMED" (as opposed to the traditional
McIDAS "NUMBERED" datasets). For a NAMED dataset, the dataset positions
returned by the server are generated by forming a list of all files
matching the DIRFILE mask (identical to a list generated by the Unix ls
command). Please note that the position of an image in a NAMED dataset
must be considered as instantaneous (rather than absolute) because the
positions are based the file list generated by the server.

Further, the files can be local or remote. Knowing the data file
name will be meaningless for remotely severed datasets.

re: file name length limit

Since the servers are written in Fortran, there is an upper limit to
the length of a file name (512 characters).

Russ Dengel

Date: Thu, 27 Jun 2002 15:26:06 -0500
From: Dave Santek <address@hidden>
Organization: Space Science & Engineering Center
To: Jianbin Yang <address@hidden>
CC: address@hidden
Subject: Re: [mdf] [Fwd: Question about using "DIRfile" in DSSERVE]

Hi Jianbin,

You may want to have more than one data set: one for each data source
[AMSU, TRMM]. Then you could access the data using the DAY= and TIME=
keywords, and not have to use position number.

dave

Date: Thu, 27 Jun 2002 16:49:06 -0400
From: "Jianbin Yang" <address@hidden>
To: address@hidden, address@hidden
Subject: Re: [mdf] [Fwd: Question about using "DIRfile" in DSSERVE]

Hi Dave,

Thanks for your help.

In fact, the files name are more longer than I mentioned in last email.
In the same Day and Time and same data source (AMSU or TRMM) there will
be more than 1 files.  For example, these files might be:
2002OLGA.WTNT42.271200.AMSU.271410.00  (the last 00 means 0-6 hours)
2002OLGA.WTNT42.271200.AMSU.271410.06 (the last 06 means 6-12 hours)
.....
2002OLGA.WTNT42.271200.AMSU.271410.24  (the last 24 means 0-24 hours)
......

Thanks.

Jianbin

NOTE: All email exchanges with Unidata User Support are recorded in the Unidata inquiry tracking system and then made publicly available through the web. If you do not want to have your interactions made available in this way, you must let us know in each email you send to us.