Back to Help on the WebTables System
Constructing Plot and Table URLs by Hand
The SLD WebTables System is mainly used to provide access to data
from the following Oracle tables:
- SLD_ONLINE.DATLOG
- SLD_ONLINE.RUNLOG
- SLD.PASS1RUN
- SLD.PASS2RUN
- SLD.OFFLDAY
- SLD.ZLOG
- SLD.JOBS
- SLD.TEST_DUCS
- SLD.RECON_RUNS
- SLD.RECON_TAPES
- SLD.OFFLINE_RUN_REPORT
- SLD.OFFLINE_DAY_REPORT
- SLD.DUCS_VERSIONS
- CATER.REPORTS
- CATER.SOLUTIONS
All of these tables currently reside on SLAC Central Unix.
Most of the elements in these tables are already available from one
of the standard interfaces.
But even if an element from one of these tables is not yet in the
interfaces, you can still access it through WebTables.
You can even access information from tables that are completely
unknown to the WebTables system.
In these cases, and for other special purposes,
you will need to construct the plot or table URL by hand.
This is best done by first using the normal interfaces to make a
plot or table that is similar to what you want,
and then modify the resulting URL to get what you really want.
Read the following explanation of the URL and then pay careful attention
to the list of additional restrictions that apply for
datanames that are not known to the WebTables system.
A typical table URL is as follows:
http://www-sld.slac.stanford.edu/htbin/wtintermed/CATER/TABLE/EXCL/
+CR.REPDATE+CR.NUM+CR.TYPE+CR.URGENCY+CR.DESCRIP
+CR.REPDATE^R^B^26-APR-97^01-MAY-97
The relevant parts of the URL are delimited by slashes.
They are as follows:
- prefix
- always "http://www-sld.slac.stanford.edu/htbin/wtintermed"
- form title
- display type
- TABLE
- SUM
- PLOT
- if PLOT is followed by ^, YMIN and YMAX plotting values are given
- invalid handling
- INCL to include rows for which some of the request data is invalid
- EXCL to exclude rows for which some of the request data is invalid
- data names
- a set of data names delimited by plus signs
(or delimited by vertical bars to indicate ratios)
- If a data name is followed by ^G, this is a "Group by" specifier.
If grouping by a date element, ^G can optionally be followed by
a binning specifier.
- ^Y indicates bin by Year
- ^M indicates bin by Month
- ^W indicates bin by Week
- ^D indicates bin by Day
- ^H indicates bin by Hour
- ^T indicates bin by Minute
- If a data name is followed by ^R, this is a range specifier.
It must then be followed by one of the following:
- ^B indicates a "Between" specifier.
The range values follow delimited by ^.
- ^L indicates a "Last" specifier.
The last count follows delimited by ^.
- ^E indicates an "Equals" specifier.
The value to match follows delimited by ^.
- ^N indicates an "Not Equals" specifier.
The value to avoid follows delimited by ^.
- ^I indicates an "Includes" specifier.
The value to search for follows delimited by ^.
- If a data name is followed by ^Q, this is a qualifier.
Qualifiers are essentially the same as range specifications,
except that the simple and advanced interfaces display them as special
options at the top of the form. Only some of the interface forms contain
any of these qualifiers. An example is the "real beams" vs
"cosmics" qualifier in the
Simple SLD Interface.
- A dataname can be a constant rather than a normal dataname.
As a silly example, if you wanted WebTables to return pi divided
by the report number, CR.NUM, the URL could contain:
Constants can be specified as integers or floating point numbers.
- Be careful not to hand-insert datanames that are from completely
unrelated tables (such as using both a dataname from the CATER tables
and a dataname from the Daily DUCS Test tables).
For such cases, the system has no intelligent way to join the tables
and will therfore do the dreaded (time-consuming and essentially useless)
"cartesian join."
The system only contains intelligent join rules for pairs of tables
that both exist in the selected "form title" (e.g., that both exist
in the SLD Main Tables interface, or that both exist in the CATER
interface).
To keep the URL from getting too long,
the datanames that are currently known to the WebTables system can have their
Oracle Owner and Table names abbreviated as follows.
- DL instead of SLD_ONLINE.DATLOG
- RL instead of SLD_ONLINE.RUNLOG
- P1 instead of SLD.PASS1RUN
- P1 instead of SLD.PASS2RUN
- OD instead of SLD.OFFLDAY
- ZL instead of SLD.ZLOG
- JB instead of SLD.JOBS
- TD instead of SLD.TEST_DUCS
- RR instead of SLD.RECON_RUNS
- RT instead of SLD.RECON_TAPES
- OC instead of SLD.OFFLINE_RUN_REPORT
- DC instead of SLD.OFFLINE_DAY_REPORT
- DA instead of SLD.DUCS_VERSIONS
- CR instead of CATER.REPORTS
- CS instead of CATER.SOLUTIONS
Datanames that are not currently known to the WebTables system
can still be accessed through WebTables.
Not only can you reference elements that are not currently known to the system,
but you can reference elements from whole tables that are not currently
known to the system.
-
Such datanames can be used as ranges as well as output columns
-
Such datanames can be used in group by columns as well as regular output columns
Since the WebTables system has less information to go on for these datanames,
the following restrictions apply:
-
If any grouping or SUM is involved,
the dataname cannot be of datatype DATE.
This restriction results from the fact that it is the glossary file that tells
the system that such a dataname is a date,
and dates require special handling in grouping and summing.
If the entire table is currently unknown to the system,
the following two additional notes apply:
-
Only a single table can be referenced (since the join statements needed
for queries of more than one table at a time are carried in the datanames.dat
files)
-
The table must be a publicly available SLD table
When you reference completely unknown tables, it doesn't matter what
form title you specify in the URL.
Just use any existing form title (e.g., SLD or CATER).
If the URL is modified to change "TABLE" into "TABLE^RAW",
the output is done in a form that is suitable for export to other applications.
-
the column headers, which normally repeat every 28 lines then appear only once
-
the final parts of the page (links, signiture and date) are omitted
You can then save the output as plain text by using your browser's
"save as...".
It only remains to cut off however much of the top of the page that you
don't want. Everything after that is straight columns of output.
You can alternately save the output by using the cut and paste features
of your desktop.
Joseph Perl
and
The Run Information Group
Last Updated: 9 October 1997