[Met_help] [rt.rap.ucar.edu #82931] History for MODE Questions
John Halley Gotway via RT
met_help at ucar.edu
Thu Nov 9 11:23:37 MST 2017
----------------------------------------------------------------
Initial Request
----------------------------------------------------------------
I am working with MODE (not MODETD yet) right now, and I have a couple of questions for you.
1.) When I select "Event Equalization", the data disappears completely and I'm not sure why.
2.) Is there a way to move the "NSTATS" selection to the bottom of the chart so that the title does not cover the number of stats? I've attached a screenshot to show how the title covers the number of stats.
3.) Jacob Carley recommended that I try to add a difference curve and use bootstrapping to show significance, and select the "show significance" so that the significant differences are in bold. I am having trouble adding a difference curve however. Could you please explain how to add the difference curve? I must be doing something wrong.
4.) Is there a way to rename a database without having to drop the database and reload all of the data again?
----------------------------------------------------------------
Complete Ticket History
----------------------------------------------------------------
Subject: MODE Questions
From: John Halley Gotway
Time: Tue Nov 07 15:23:23 2017
Tracey,
I turned your questions into a met-help ticket and am answering
through met-help.
Let me answer the questions that I can...
(4) Unfortunately, as far as I can tell, there's no simple command you
can use to rename a database. However, I did follow some instructions
recommended here:
http://www.rndblog.com/how-to-rename-a-database-in-mysql/
# Dump existing database to a sql file
mysqldump -h{hostname} -u{username} -p{password} mv_{dbname} >
mv_{dbname}.sql
# Create a new database name
mysql -h{hostname} -u{username} -p{password} -e "create database
mv_{new_dbname}"
# Populate new database
mysql -h{hostname} -u{username} -p{password} mv_{new_dbname} <
mv_{dbname}.sql
# Delete old database
mysql -h{hostname} -u{username} -p{password} -e "delete database
mv_{dbname}"
Seems like that should be easier!
(2) There is no easy way to move the "NSTATS" selection to the bottom
of the chart. But you have some other options...
- Your title uses 4 lines. You could squeeze more info onto 3 lines
instead.
- You could move some info from the "Title" to the "Caption" which
shows up at the bottom-left corner of the plot.
- You could adjust the margins of the plot to give yourself more
real-estate above the plot. On the "Formatting" Tab, look in the
"Plot Formatting" box at the "Margins". There are 4 values listed
corresponding to the margins at the bottom, left, top, and right sides
of the plot. Increase the value in the 3rd box to give yourself more
room above the plot.
Hopefully one of those will do the trick!
I think I'll need to look more closely at (1) and (2) with Tatiana.
Thanks,
John
------------------------------------------------
Subject: MODE Questions
From: tracey.dorian at noaa.gov
Time: Wed Nov 08 08:34:50 2017
Hi John,
Thanks for forwarding the question to MET help. How can I submit
questions
in the future so that they go through MET-help?
Thank you for looking into instructions on how to change database
name. I
will probably keep the same name this time, but in the future it's
good to
know there is a way to change the name.
I changed the title to only two lines, and changed the "Perpendicular
offset" to -0.6 in the "Title" formatting under the "Formatting" tab.
This
seemed to solve the issue of NSTATS being blocked (see attached).
Are you able to add a difference curve? When I choose "Add derived
curve",
I don't get any error or anything, but nothing is plotted.
Thanks for your help!
Tracey
On Tue, Nov 7, 2017 at 5:23 PM, John Halley Gotway via RT
<met_help at ucar.edu
> wrote:
> Tracey,
>
> I turned your questions into a met-help ticket and am answering
through
> met-help.
>
> Let me answer the questions that I can...
>
> (4) Unfortunately, as far as I can tell, there's no simple command
you can
> use to rename a database. However, I did follow some instructions
> recommended here:
> http://www.rndblog.com/how-to-rename-a-database-in-mysql/
>
> # Dump existing database to a sql file
> mysqldump -h{hostname} -u{username} -p{password} mv_{dbname} >
> mv_{dbname}.sql
>
> # Create a new database name
> mysql -h{hostname} -u{username} -p{password} -e "create database
> mv_{new_dbname}"
>
> # Populate new database
> mysql -h{hostname} -u{username} -p{password} mv_{new_dbname} <
> mv_{dbname}.sql
>
> # Delete old database
> mysql -h{hostname} -u{username} -p{password} -e "delete database
> mv_{dbname}"
>
> Seems like that should be easier!
>
> (2) There is no easy way to move the "NSTATS" selection to the
bottom of
> the chart. But you have some other options...
> - Your title uses 4 lines. You could squeeze more info onto 3
lines
> instead.
> - You could move some info from the "Title" to the "Caption" which
shows
> up at the bottom-left corner of the plot.
> - You could adjust the margins of the plot to give yourself more
> real-estate above the plot. On the "Formatting" Tab, look in the
"Plot
> Formatting" box at the "Margins". There are 4 values listed
corresponding
> to the margins at the bottom, left, top, and right sides of the
plot.
> Increase the value in the 3rd box to give yourself more room above
the plot.
>
> Hopefully one of those will do the trick!
>
> I think I'll need to look more closely at (1) and (2) with Tatiana.
>
> Thanks,
> John
>
>
>
>
--
Tracey Dorian
NOAA/NCEP/EMC
IMSG contractor
5830 University Research Court
College Park, MD 20740
------------------------------------------------
Subject: MODE Questions
From: John Halley Gotway
Time: Wed Nov 08 13:38:39 2017
Adding the 2 attachments that Tracey sent in her original email.
------------------------------------------------
Subject: MODE Questions
From: John Halley Gotway
Time: Wed Nov 08 13:55:18 2017
Hi Tracey,
You can just send future questions directly to "met_help at ucar.edu".
Tatiana pointed out to me that my instructions for renaming a database
were
slightly off. When you run "mysqldump" you should use some command
line
options to also dump the stored procedures. She wasn't sure on the
exact
options, but I'm guessing it's:
-R, --routines Dump stored routines (functions and
procedures).
So don't delete your old database before testing to make sure the new
one
works well!
I logged onto the NCEP VPN and loaded your XML. I did notice that the
XML
upload had a small problem. It didn't add entries in the Series
Formatting
section for the series that were defined in the XML. I talked to
Tatiana
about that and she's taking a look.
As for adding a difference curve, that logic is not supported for MODE
data. It's supported for STAT data, but not MODE. So that'd require
enhancements to METViewer to support.
Thanks,
John
On Wed, Nov 8, 2017 at 8:34 AM, tracey.dorian at noaa.gov via RT <
met_help at ucar.edu> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=82931 >
>
> Hi John,
>
> Thanks for forwarding the question to MET help. How can I submit
questions
> in the future so that they go through MET-help?
>
> Thank you for looking into instructions on how to change database
name. I
> will probably keep the same name this time, but in the future it's
good to
> know there is a way to change the name.
>
> I changed the title to only two lines, and changed the
"Perpendicular
> offset" to -0.6 in the "Title" formatting under the "Formatting"
tab. This
> seemed to solve the issue of NSTATS being blocked (see attached).
>
> Are you able to add a difference curve? When I choose "Add derived
curve",
> I don't get any error or anything, but nothing is plotted.
>
> Thanks for your help!
> Tracey
>
>
>
>
>
> On Tue, Nov 7, 2017 at 5:23 PM, John Halley Gotway via RT <
> met_help at ucar.edu
> > wrote:
>
> > Tracey,
> >
> > I turned your questions into a met-help ticket and am answering
through
> > met-help.
> >
> > Let me answer the questions that I can...
> >
> > (4) Unfortunately, as far as I can tell, there's no simple command
you
> can
> > use to rename a database. However, I did follow some instructions
> > recommended here:
> > http://www.rndblog.com/how-to-rename-a-database-in-mysql/
> >
> > # Dump existing database to a sql file
> > mysqldump -h{hostname} -u{username} -p{password} mv_{dbname} >
> > mv_{dbname}.sql
> >
> > # Create a new database name
> > mysql -h{hostname} -u{username} -p{password} -e "create database
> > mv_{new_dbname}"
> >
> > # Populate new database
> > mysql -h{hostname} -u{username} -p{password} mv_{new_dbname} <
> > mv_{dbname}.sql
> >
> > # Delete old database
> > mysql -h{hostname} -u{username} -p{password} -e "delete database
> > mv_{dbname}"
> >
> > Seems like that should be easier!
> >
> > (2) There is no easy way to move the "NSTATS" selection to the
bottom of
> > the chart. But you have some other options...
> > - Your title uses 4 lines. You could squeeze more info onto 3
lines
> > instead.
> > - You could move some info from the "Title" to the "Caption"
which
> shows
> > up at the bottom-left corner of the plot.
> > - You could adjust the margins of the plot to give yourself more
> > real-estate above the plot. On the "Formatting" Tab, look in the
"Plot
> > Formatting" box at the "Margins". There are 4 values listed
> corresponding
> > to the margins at the bottom, left, top, and right sides of the
plot.
> > Increase the value in the 3rd box to give yourself more room above
the
> plot.
> >
> > Hopefully one of those will do the trick!
> >
> > I think I'll need to look more closely at (1) and (2) with
Tatiana.
> >
> > Thanks,
> > John
> >
> >
> >
> >
>
>
> --
> Tracey Dorian
> NOAA/NCEP/EMC
> IMSG contractor
> 5830 University Research Court
> College Park, MD 20740
>
>
------------------------------------------------
Subject: MODE Questions
From: John Halley Gotway
Time: Wed Nov 08 14:07:40 2017
Tracey,
As for event equalization, I did get that to work well. Just deselect
the
checkbox for "Equalize" next to the FCST_INIT list of values in the
"Fixed
Values" section. And then make your plot again.
This is a pretty subtle point.
To turn ON event equalization, you click the "Event Equalization"
checkbox
at the bottom of the fixed values section. There are also many
smaller
checkboxes labelled "Equalize" next to each entry in the "Fixed
Values" and
"Independent Variables" section. When you select the master "Event
Equalization" checkbox, all of the "Equalize" checkboxes are also
selected
by default.
What does this all mean? In the simplest form, event equalization is
just
comparing 2 or more lists of dates and picking out the dates that are
common to all the lists. The tricky part is deciding *how* to define
those
lists of dates in the first place! The "Equalize" checkboxes tell
METViewer which fields to include when defining the lists of dates to
be
equalized. It makes sense to include the FCST_RAD and FCST_THRESH
because
you only want to use dates where you have output for all the
combinations
of those. But we do NOT want to include FCST_INIT in case definition
because then we'd end up with a bunch of lists of length 1 that have
no
dates in common!
To put it simply, whenever you include a list of dates (FCST_INIT or
FCST_VALID) in the fixed values section, you don't want to use them in
the
equalization logic. If you make a plot where event equalization
removes
everything, look closely at the "Equalize" checkboxes you have
selected.
Thanks,
John
On Wed, Nov 8, 2017 at 1:54 PM, John Halley Gotway <johnhg at ucar.edu>
wrote:
> Hi Tracey,
>
> You can just send future questions directly to "met_help at ucar.edu".
>
> Tatiana pointed out to me that my instructions for renaming a
database
> were slightly off. When you run "mysqldump" you should use some
command
> line options to also dump the stored procedures. She wasn't sure on
the
> exact options, but I'm guessing it's:
> -R, --routines Dump stored routines (functions and
procedures).
>
> So don't delete your old database before testing to make sure the
new one
> works well!
>
> I logged onto the NCEP VPN and loaded your XML. I did notice that
the XML
> upload had a small problem. It didn't add entries in the Series
Formatting
> section for the series that were defined in the XML. I talked to
Tatiana
> about that and she's taking a look.
>
> As for adding a difference curve, that logic is not supported for
MODE
> data. It's supported for STAT data, but not MODE. So that'd
require
> enhancements to METViewer to support.
>
> Thanks,
> John
>
> On Wed, Nov 8, 2017 at 8:34 AM, tracey.dorian at noaa.gov via RT <
> met_help at ucar.edu> wrote:
>
>>
>> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=82931 >
>>
>> Hi John,
>>
>> Thanks for forwarding the question to MET help. How can I submit
>> questions
>> in the future so that they go through MET-help?
>>
>> Thank you for looking into instructions on how to change database
name. I
>> will probably keep the same name this time, but in the future it's
good to
>> know there is a way to change the name.
>>
>> I changed the title to only two lines, and changed the
"Perpendicular
>> offset" to -0.6 in the "Title" formatting under the "Formatting"
tab.
>> This
>> seemed to solve the issue of NSTATS being blocked (see attached).
>>
>> Are you able to add a difference curve? When I choose "Add derived
>> curve",
>> I don't get any error or anything, but nothing is plotted.
>>
>> Thanks for your help!
>> Tracey
>>
>>
>>
>>
>>
>> On Tue, Nov 7, 2017 at 5:23 PM, John Halley Gotway via RT <
>> met_help at ucar.edu
>> > wrote:
>>
>> > Tracey,
>> >
>> > I turned your questions into a met-help ticket and am answering
through
>> > met-help.
>> >
>> > Let me answer the questions that I can...
>> >
>> > (4) Unfortunately, as far as I can tell, there's no simple
command you
>> can
>> > use to rename a database. However, I did follow some
instructions
>> > recommended here:
>> > http://www.rndblog.com/how-to-rename-a-database-in-mysql/
>> >
>> > # Dump existing database to a sql file
>> > mysqldump -h{hostname} -u{username} -p{password} mv_{dbname} >
>> > mv_{dbname}.sql
>> >
>> > # Create a new database name
>> > mysql -h{hostname} -u{username} -p{password} -e "create database
>> > mv_{new_dbname}"
>> >
>> > # Populate new database
>> > mysql -h{hostname} -u{username} -p{password} mv_{new_dbname} <
>> > mv_{dbname}.sql
>> >
>> > # Delete old database
>> > mysql -h{hostname} -u{username} -p{password} -e "delete database
>> > mv_{dbname}"
>> >
>> > Seems like that should be easier!
>> >
>> > (2) There is no easy way to move the "NSTATS" selection to the
bottom of
>> > the chart. But you have some other options...
>> > - Your title uses 4 lines. You could squeeze more info onto 3
lines
>> > instead.
>> > - You could move some info from the "Title" to the "Caption"
which
>> shows
>> > up at the bottom-left corner of the plot.
>> > - You could adjust the margins of the plot to give yourself
more
>> > real-estate above the plot. On the "Formatting" Tab, look in the
"Plot
>> > Formatting" box at the "Margins". There are 4 values listed
>> corresponding
>> > to the margins at the bottom, left, top, and right sides of the
plot.
>> > Increase the value in the 3rd box to give yourself more room
above the
>> plot.
>> >
>> > Hopefully one of those will do the trick!
>> >
>> > I think I'll need to look more closely at (1) and (2) with
Tatiana.
>> >
>> > Thanks,
>> > John
>> >
>> >
>> >
>> >
>>
>>
>> --
>> Tracey Dorian
>> NOAA/NCEP/EMC
>> IMSG contractor
>> 5830 University Research Court
>> College Park, MD 20740
>>
>>
>
------------------------------------------------
Subject: MODE Questions
From: tracey.dorian at noaa.gov
Time: Wed Nov 08 16:03:20 2017
Hi John,
Thanks for much for your clear explanations. I unchecked the "Event
Equalization" for the "FCST_INIT" and the event equalization worked.
I did
not know that difference curves were not available for MODE, but I
guess
I'm glad I wasn't doing anything wrong! Are there any plans to
include
this in future versions? Would it be something that is easy to add?
I think my last question from my original email was regarding the
bootstrapping confidence intervals for statistical significance. When
I
choose "boot" under "Conf Interval" at the bottom in the "Series
Formatting", I don't see any bars around the data, and so nothing is
bolded
when I select "Show Significance." Do you recommend that I just stick
with
"Conf Interval" = "std"?
Thanks again!
Tracey
On Wed, Nov 8, 2017 at 4:07 PM, John Halley Gotway via RT
<met_help at ucar.edu
> wrote:
> Tracey,
>
> As for event equalization, I did get that to work well. Just
deselect the
> checkbox for "Equalize" next to the FCST_INIT list of values in the
"Fixed
> Values" section. And then make your plot again.
>
> This is a pretty subtle point.
>
> To turn ON event equalization, you click the "Event Equalization"
checkbox
> at the bottom of the fixed values section. There are also many
smaller
> checkboxes labelled "Equalize" next to each entry in the "Fixed
Values" and
> "Independent Variables" section. When you select the master "Event
> Equalization" checkbox, all of the "Equalize" checkboxes are also
selected
> by default.
>
> What does this all mean? In the simplest form, event equalization
is just
> comparing 2 or more lists of dates and picking out the dates that
are
> common to all the lists. The tricky part is deciding *how* to
define those
> lists of dates in the first place! The "Equalize" checkboxes tell
> METViewer which fields to include when defining the lists of dates
to be
> equalized. It makes sense to include the FCST_RAD and FCST_THRESH
because
> you only want to use dates where you have output for all the
combinations
> of those. But we do NOT want to include FCST_INIT in case
definition
> because then we'd end up with a bunch of lists of length 1 that have
no
> dates in common!
>
> To put it simply, whenever you include a list of dates (FCST_INIT or
> FCST_VALID) in the fixed values section, you don't want to use them
in the
> equalization logic. If you make a plot where event equalization
removes
> everything, look closely at the "Equalize" checkboxes you have
selected.
>
> Thanks,
> John
>
>
>
> On Wed, Nov 8, 2017 at 1:54 PM, John Halley Gotway <johnhg at ucar.edu>
> wrote:
>
> > Hi Tracey,
> >
> > You can just send future questions directly to
"met_help at ucar.edu".
> >
> > Tatiana pointed out to me that my instructions for renaming a
database
> > were slightly off. When you run "mysqldump" you should use some
command
> > line options to also dump the stored procedures. She wasn't sure
on the
> > exact options, but I'm guessing it's:
> > -R, --routines Dump stored routines (functions and
procedures).
> >
> > So don't delete your old database before testing to make sure the
new one
> > works well!
> >
> > I logged onto the NCEP VPN and loaded your XML. I did notice that
the
> XML
> > upload had a small problem. It didn't add entries in the Series
> Formatting
> > section for the series that were defined in the XML. I talked to
Tatiana
> > about that and she's taking a look.
> >
> > As for adding a difference curve, that logic is not supported for
MODE
> > data. It's supported for STAT data, but not MODE. So that'd
require
> > enhancements to METViewer to support.
> >
> > Thanks,
> > John
> >
> > On Wed, Nov 8, 2017 at 8:34 AM, tracey.dorian at noaa.gov via RT <
> > met_help at ucar.edu> wrote:
> >
> >>
> >> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=82931 >
> >>
> >> Hi John,
> >>
> >> Thanks for forwarding the question to MET help. How can I submit
> >> questions
> >> in the future so that they go through MET-help?
> >>
> >> Thank you for looking into instructions on how to change database
> name. I
> >> will probably keep the same name this time, but in the future
it's good
> to
> >> know there is a way to change the name.
> >>
> >> I changed the title to only two lines, and changed the
"Perpendicular
> >> offset" to -0.6 in the "Title" formatting under the "Formatting"
tab.
> >> This
> >> seemed to solve the issue of NSTATS being blocked (see attached).
> >>
> >> Are you able to add a difference curve? When I choose "Add
derived
> >> curve",
> >> I don't get any error or anything, but nothing is plotted.
> >>
> >> Thanks for your help!
> >> Tracey
> >>
> >>
> >>
> >>
> >>
> >> On Tue, Nov 7, 2017 at 5:23 PM, John Halley Gotway via RT <
> >> met_help at ucar.edu
> >> > wrote:
> >>
> >> > Tracey,
> >> >
> >> > I turned your questions into a met-help ticket and am answering
> through
> >> > met-help.
> >> >
> >> > Let me answer the questions that I can...
> >> >
> >> > (4) Unfortunately, as far as I can tell, there's no simple
command you
> >> can
> >> > use to rename a database. However, I did follow some
instructions
> >> > recommended here:
> >> > http://www.rndblog.com/how-to-rename-a-database-in-mysql/
> >> >
> >> > # Dump existing database to a sql file
> >> > mysqldump -h{hostname} -u{username} -p{password} mv_{dbname} >
> >> > mv_{dbname}.sql
> >> >
> >> > # Create a new database name
> >> > mysql -h{hostname} -u{username} -p{password} -e "create
database
> >> > mv_{new_dbname}"
> >> >
> >> > # Populate new database
> >> > mysql -h{hostname} -u{username} -p{password} mv_{new_dbname} <
> >> > mv_{dbname}.sql
> >> >
> >> > # Delete old database
> >> > mysql -h{hostname} -u{username} -p{password} -e "delete
database
> >> > mv_{dbname}"
> >> >
> >> > Seems like that should be easier!
> >> >
> >> > (2) There is no easy way to move the "NSTATS" selection to the
bottom
> of
> >> > the chart. But you have some other options...
> >> > - Your title uses 4 lines. You could squeeze more info onto
3 lines
> >> > instead.
> >> > - You could move some info from the "Title" to the "Caption"
which
> >> shows
> >> > up at the bottom-left corner of the plot.
> >> > - You could adjust the margins of the plot to give yourself
more
> >> > real-estate above the plot. On the "Formatting" Tab, look in
the
> "Plot
> >> > Formatting" box at the "Margins". There are 4 values listed
> >> corresponding
> >> > to the margins at the bottom, left, top, and right sides of the
plot.
> >> > Increase the value in the 3rd box to give yourself more room
above the
> >> plot.
> >> >
> >> > Hopefully one of those will do the trick!
> >> >
> >> > I think I'll need to look more closely at (1) and (2) with
Tatiana.
> >> >
> >> > Thanks,
> >> > John
> >> >
> >> >
> >> >
> >> >
> >>
> >>
> >> --
> >> Tracey Dorian
> >> NOAA/NCEP/EMC
> >> IMSG contractor
> >> 5830 University Research Court
> >> College Park, MD 20740
> >>
> >>
> >
>
>
--
Tracey Dorian
NOAA/NCEP/EMC
IMSG contractor
5830 University Research Court
College Park, MD 20740
------------------------------------------------
Subject: MODE Questions
From: John Halley Gotway
Time: Wed Nov 08 19:09:05 2017
Tracey,
Bootstrapping applies when you’re “aggregating” your data. For
example,
when you aggregate multiple SL1L2 lines together into a single one.
There
is no way to aggregate MODE object attributes. Instead you just plot
the
mean or median of a distribution of values. And when you’re plotting
the
mean or median then the std error is the appropriate CI method.
So bootstrapping doesn’t really apply to MODE output.
Technically, you “could” bootstrap the computation of the mean or
median...
but the statisticians say that’s overkill.
As for computing differences for MODE output, I’ll leave that to
Tatiana to
answer. The MODE output is a lot messier than the stat output. So
we’d
need to think carefully about the appropriate way of computing
differences.
Tressa Fowler will be visiting NCEP for a few weeks. Maybe the two of
you
could discuss how the logic should work?
John
On Wed, Nov 8, 2017 at 4:03 PM tracey.dorian at noaa.gov via RT <
met_help at ucar.edu> wrote:
>
> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=82931 >
>
> Hi John,
>
> Thanks for much for your clear explanations. I unchecked the "Event
> Equalization" for the "FCST_INIT" and the event equalization worked.
I did
> not know that difference curves were not available for MODE, but I
guess
> I'm glad I wasn't doing anything wrong! Are there any plans to
include
> this in future versions? Would it be something that is easy to add?
>
> I think my last question from my original email was regarding the
> bootstrapping confidence intervals for statistical significance.
When I
> choose "boot" under "Conf Interval" at the bottom in the "Series
> Formatting", I don't see any bars around the data, and so nothing is
bolded
> when I select "Show Significance." Do you recommend that I just
stick with
> "Conf Interval" = "std"?
>
> Thanks again!
> Tracey
>
> On Wed, Nov 8, 2017 at 4:07 PM, John Halley Gotway via RT <
> met_help at ucar.edu
> > wrote:
>
> > Tracey,
> >
> > As for event equalization, I did get that to work well. Just
deselect
> the
> > checkbox for "Equalize" next to the FCST_INIT list of values in
the
> "Fixed
> > Values" section. And then make your plot again.
> >
> > This is a pretty subtle point.
> >
> > To turn ON event equalization, you click the "Event Equalization"
> checkbox
> > at the bottom of the fixed values section. There are also many
smaller
> > checkboxes labelled "Equalize" next to each entry in the "Fixed
Values"
> and
> > "Independent Variables" section. When you select the master
"Event
> > Equalization" checkbox, all of the "Equalize" checkboxes are also
> selected
> > by default.
> >
> > What does this all mean? In the simplest form, event equalization
is
> just
> > comparing 2 or more lists of dates and picking out the dates that
are
> > common to all the lists. The tricky part is deciding *how* to
define
> those
> > lists of dates in the first place! The "Equalize" checkboxes tell
> > METViewer which fields to include when defining the lists of dates
to be
> > equalized. It makes sense to include the FCST_RAD and FCST_THRESH
> because
> > you only want to use dates where you have output for all the
combinations
> > of those. But we do NOT want to include FCST_INIT in case
definition
> > because then we'd end up with a bunch of lists of length 1 that
have no
> > dates in common!
> >
> > To put it simply, whenever you include a list of dates (FCST_INIT
or
> > FCST_VALID) in the fixed values section, you don't want to use
them in
> the
> > equalization logic. If you make a plot where event equalization
removes
> > everything, look closely at the "Equalize" checkboxes you have
selected.
> >
> > Thanks,
> > John
> >
> >
> >
> > On Wed, Nov 8, 2017 at 1:54 PM, John Halley Gotway
<johnhg at ucar.edu>
> > wrote:
> >
> > > Hi Tracey,
> > >
> > > You can just send future questions directly to
"met_help at ucar.edu".
> > >
> > > Tatiana pointed out to me that my instructions for renaming a
database
> > > were slightly off. When you run "mysqldump" you should use some
> command
> > > line options to also dump the stored procedures. She wasn't
sure on
> the
> > > exact options, but I'm guessing it's:
> > > -R, --routines Dump stored routines (functions and
procedures).
> > >
> > > So don't delete your old database before testing to make sure
the new
> one
> > > works well!
> > >
> > > I logged onto the NCEP VPN and loaded your XML. I did notice
that the
> > XML
> > > upload had a small problem. It didn't add entries in the Series
> > Formatting
> > > section for the series that were defined in the XML. I talked
to
> Tatiana
> > > about that and she's taking a look.
> > >
> > > As for adding a difference curve, that logic is not supported
for MODE
> > > data. It's supported for STAT data, but not MODE. So that'd
require
> > > enhancements to METViewer to support.
> > >
> > > Thanks,
> > > John
> > >
> > > On Wed, Nov 8, 2017 at 8:34 AM, tracey.dorian at noaa.gov via RT <
> > > met_help at ucar.edu> wrote:
> > >
> > >>
> > >> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=82931 >
> > >>
> > >> Hi John,
> > >>
> > >> Thanks for forwarding the question to MET help. How can I
submit
> > >> questions
> > >> in the future so that they go through MET-help?
> > >>
> > >> Thank you for looking into instructions on how to change
database
> > name. I
> > >> will probably keep the same name this time, but in the future
it's
> good
> > to
> > >> know there is a way to change the name.
> > >>
> > >> I changed the title to only two lines, and changed the
"Perpendicular
> > >> offset" to -0.6 in the "Title" formatting under the
"Formatting" tab.
> > >> This
> > >> seemed to solve the issue of NSTATS being blocked (see
attached).
> > >>
> > >> Are you able to add a difference curve? When I choose "Add
derived
> > >> curve",
> > >> I don't get any error or anything, but nothing is plotted.
> > >>
> > >> Thanks for your help!
> > >> Tracey
> > >>
> > >>
> > >>
> > >>
> > >>
> > >> On Tue, Nov 7, 2017 at 5:23 PM, John Halley Gotway via RT <
> > >> met_help at ucar.edu
> > >> > wrote:
> > >>
> > >> > Tracey,
> > >> >
> > >> > I turned your questions into a met-help ticket and am
answering
> > through
> > >> > met-help.
> > >> >
> > >> > Let me answer the questions that I can...
> > >> >
> > >> > (4) Unfortunately, as far as I can tell, there's no simple
command
> you
> > >> can
> > >> > use to rename a database. However, I did follow some
instructions
> > >> > recommended here:
> > >> > http://www.rndblog.com/how-to-rename-a-database-in-mysql/
> > >> >
> > >> > # Dump existing database to a sql file
> > >> > mysqldump -h{hostname} -u{username} -p{password} mv_{dbname}
>
> > >> > mv_{dbname}.sql
> > >> >
> > >> > # Create a new database name
> > >> > mysql -h{hostname} -u{username} -p{password} -e "create
database
> > >> > mv_{new_dbname}"
> > >> >
> > >> > # Populate new database
> > >> > mysql -h{hostname} -u{username} -p{password} mv_{new_dbname}
<
> > >> > mv_{dbname}.sql
> > >> >
> > >> > # Delete old database
> > >> > mysql -h{hostname} -u{username} -p{password} -e "delete
database
> > >> > mv_{dbname}"
> > >> >
> > >> > Seems like that should be easier!
> > >> >
> > >> > (2) There is no easy way to move the "NSTATS" selection to
the
> bottom
> > of
> > >> > the chart. But you have some other options...
> > >> > - Your title uses 4 lines. You could squeeze more info
onto 3
> lines
> > >> > instead.
> > >> > - You could move some info from the "Title" to the
"Caption" which
> > >> shows
> > >> > up at the bottom-left corner of the plot.
> > >> > - You could adjust the margins of the plot to give yourself
more
> > >> > real-estate above the plot. On the "Formatting" Tab, look in
the
> > "Plot
> > >> > Formatting" box at the "Margins". There are 4 values listed
> > >> corresponding
> > >> > to the margins at the bottom, left, top, and right sides of
the
> plot.
> > >> > Increase the value in the 3rd box to give yourself more room
above
> the
> > >> plot.
> > >> >
> > >> > Hopefully one of those will do the trick!
> > >> >
> > >> > I think I'll need to look more closely at (1) and (2) with
Tatiana.
> > >> >
> > >> > Thanks,
> > >> > John
> > >> >
> > >> >
> > >> >
> > >> >
> > >>
> > >>
> > >> --
> > >> Tracey Dorian
> > >> NOAA/NCEP/EMC
> > >> IMSG contractor
> > >> 5830 University Research Court
> > >> College Park, MD 20740
> > >>
> > >>
> > >
> >
> >
>
>
> --
> Tracey Dorian
> NOAA/NCEP/EMC
> IMSG contractor
> 5830 University Research Court
> College Park, MD 20740
>
>
------------------------------------------------
Subject: MODE Questions
From: tracey.dorian at noaa.gov
Time: Thu Nov 09 08:08:41 2017
Hi John,
Sure, that sounds good. When will Tressa be visiting?
Thanks again for your assistance,
Tracey
On Wed, Nov 8, 2017 at 9:09 PM, John Halley Gotway via RT
<met_help at ucar.edu
> wrote:
> Tracey,
>
> Bootstrapping applies when you’re “aggregating” your data. For
example,
> when you aggregate multiple SL1L2 lines together into a single one.
There
> is no way to aggregate MODE object attributes. Instead you just
plot the
> mean or median of a distribution of values. And when you’re
plotting the
> mean or median then the std error is the appropriate CI method.
>
> So bootstrapping doesn’t really apply to MODE output.
>
> Technically, you “could” bootstrap the computation of the mean or
median...
> but the statisticians say that’s overkill.
>
> As for computing differences for MODE output, I’ll leave that to
Tatiana to
> answer. The MODE output is a lot messier than the stat output. So
we’d
> need to think carefully about the appropriate way of computing
differences.
>
> Tressa Fowler will be visiting NCEP for a few weeks. Maybe the two
of you
> could discuss how the logic should work?
>
> John
>
> On Wed, Nov 8, 2017 at 4:03 PM tracey.dorian at noaa.gov via RT <
> met_help at ucar.edu> wrote:
>
> >
> > <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=82931 >
> >
> > Hi John,
> >
> > Thanks for much for your clear explanations. I unchecked the
"Event
> > Equalization" for the "FCST_INIT" and the event equalization
worked. I
> did
> > not know that difference curves were not available for MODE, but I
guess
> > I'm glad I wasn't doing anything wrong! Are there any plans to
include
> > this in future versions? Would it be something that is easy to
add?
> >
> > I think my last question from my original email was regarding the
> > bootstrapping confidence intervals for statistical significance.
When I
> > choose "boot" under "Conf Interval" at the bottom in the "Series
> > Formatting", I don't see any bars around the data, and so nothing
is
> bolded
> > when I select "Show Significance." Do you recommend that I just
stick
> with
> > "Conf Interval" = "std"?
> >
> > Thanks again!
> > Tracey
> >
> > On Wed, Nov 8, 2017 at 4:07 PM, John Halley Gotway via RT <
> > met_help at ucar.edu
> > > wrote:
> >
> > > Tracey,
> > >
> > > As for event equalization, I did get that to work well. Just
deselect
> > the
> > > checkbox for "Equalize" next to the FCST_INIT list of values in
the
> > "Fixed
> > > Values" section. And then make your plot again.
> > >
> > > This is a pretty subtle point.
> > >
> > > To turn ON event equalization, you click the "Event
Equalization"
> > checkbox
> > > at the bottom of the fixed values section. There are also many
smaller
> > > checkboxes labelled "Equalize" next to each entry in the "Fixed
Values"
> > and
> > > "Independent Variables" section. When you select the master
"Event
> > > Equalization" checkbox, all of the "Equalize" checkboxes are
also
> > selected
> > > by default.
> > >
> > > What does this all mean? In the simplest form, event
equalization is
> > just
> > > comparing 2 or more lists of dates and picking out the dates
that are
> > > common to all the lists. The tricky part is deciding *how* to
define
> > those
> > > lists of dates in the first place! The "Equalize" checkboxes
tell
> > > METViewer which fields to include when defining the lists of
dates to
> be
> > > equalized. It makes sense to include the FCST_RAD and
FCST_THRESH
> > because
> > > you only want to use dates where you have output for all the
> combinations
> > > of those. But we do NOT want to include FCST_INIT in case
definition
> > > because then we'd end up with a bunch of lists of length 1 that
have no
> > > dates in common!
> > >
> > > To put it simply, whenever you include a list of dates
(FCST_INIT or
> > > FCST_VALID) in the fixed values section, you don't want to use
them in
> > the
> > > equalization logic. If you make a plot where event equalization
> removes
> > > everything, look closely at the "Equalize" checkboxes you have
> selected.
> > >
> > > Thanks,
> > > John
> > >
> > >
> > >
> > > On Wed, Nov 8, 2017 at 1:54 PM, John Halley Gotway
<johnhg at ucar.edu>
> > > wrote:
> > >
> > > > Hi Tracey,
> > > >
> > > > You can just send future questions directly to
"met_help at ucar.edu".
> > > >
> > > > Tatiana pointed out to me that my instructions for renaming a
> database
> > > > were slightly off. When you run "mysqldump" you should use
some
> > command
> > > > line options to also dump the stored procedures. She wasn't
sure on
> > the
> > > > exact options, but I'm guessing it's:
> > > > -R, --routines Dump stored routines (functions and
> procedures).
> > > >
> > > > So don't delete your old database before testing to make sure
the new
> > one
> > > > works well!
> > > >
> > > > I logged onto the NCEP VPN and loaded your XML. I did notice
that
> the
> > > XML
> > > > upload had a small problem. It didn't add entries in the
Series
> > > Formatting
> > > > section for the series that were defined in the XML. I talked
to
> > Tatiana
> > > > about that and she's taking a look.
> > > >
> > > > As for adding a difference curve, that logic is not supported
for
> MODE
> > > > data. It's supported for STAT data, but not MODE. So that'd
require
> > > > enhancements to METViewer to support.
> > > >
> > > > Thanks,
> > > > John
> > > >
> > > > On Wed, Nov 8, 2017 at 8:34 AM, tracey.dorian at noaa.gov via RT
<
> > > > met_help at ucar.edu> wrote:
> > > >
> > > >>
> > > >> <URL: https://rt.rap.ucar.edu/rt/Ticket/Display.html?id=82931
>
> > > >>
> > > >> Hi John,
> > > >>
> > > >> Thanks for forwarding the question to MET help. How can I
submit
> > > >> questions
> > > >> in the future so that they go through MET-help?
> > > >>
> > > >> Thank you for looking into instructions on how to change
database
> > > name. I
> > > >> will probably keep the same name this time, but in the future
it's
> > good
> > > to
> > > >> know there is a way to change the name.
> > > >>
> > > >> I changed the title to only two lines, and changed the
> "Perpendicular
> > > >> offset" to -0.6 in the "Title" formatting under the
"Formatting"
> tab.
> > > >> This
> > > >> seemed to solve the issue of NSTATS being blocked (see
attached).
> > > >>
> > > >> Are you able to add a difference curve? When I choose "Add
derived
> > > >> curve",
> > > >> I don't get any error or anything, but nothing is plotted.
> > > >>
> > > >> Thanks for your help!
> > > >> Tracey
> > > >>
> > > >>
> > > >>
> > > >>
> > > >>
> > > >> On Tue, Nov 7, 2017 at 5:23 PM, John Halley Gotway via RT <
> > > >> met_help at ucar.edu
> > > >> > wrote:
> > > >>
> > > >> > Tracey,
> > > >> >
> > > >> > I turned your questions into a met-help ticket and am
answering
> > > through
> > > >> > met-help.
> > > >> >
> > > >> > Let me answer the questions that I can...
> > > >> >
> > > >> > (4) Unfortunately, as far as I can tell, there's no simple
command
> > you
> > > >> can
> > > >> > use to rename a database. However, I did follow some
instructions
> > > >> > recommended here:
> > > >> > http://www.rndblog.com/how-to-rename-a-database-in-
mysql/
> > > >> >
> > > >> > # Dump existing database to a sql file
> > > >> > mysqldump -h{hostname} -u{username} -p{password}
mv_{dbname} >
> > > >> > mv_{dbname}.sql
> > > >> >
> > > >> > # Create a new database name
> > > >> > mysql -h{hostname} -u{username} -p{password} -e "create
database
> > > >> > mv_{new_dbname}"
> > > >> >
> > > >> > # Populate new database
> > > >> > mysql -h{hostname} -u{username} -p{password}
mv_{new_dbname} <
> > > >> > mv_{dbname}.sql
> > > >> >
> > > >> > # Delete old database
> > > >> > mysql -h{hostname} -u{username} -p{password} -e "delete
database
> > > >> > mv_{dbname}"
> > > >> >
> > > >> > Seems like that should be easier!
> > > >> >
> > > >> > (2) There is no easy way to move the "NSTATS" selection to
the
> > bottom
> > > of
> > > >> > the chart. But you have some other options...
> > > >> > - Your title uses 4 lines. You could squeeze more info
onto 3
> > lines
> > > >> > instead.
> > > >> > - You could move some info from the "Title" to the
"Caption"
> which
> > > >> shows
> > > >> > up at the bottom-left corner of the plot.
> > > >> > - You could adjust the margins of the plot to give
yourself more
> > > >> > real-estate above the plot. On the "Formatting" Tab, look
in the
> > > "Plot
> > > >> > Formatting" box at the "Margins". There are 4 values
listed
> > > >> corresponding
> > > >> > to the margins at the bottom, left, top, and right sides of
the
> > plot.
> > > >> > Increase the value in the 3rd box to give yourself more
room above
> > the
> > > >> plot.
> > > >> >
> > > >> > Hopefully one of those will do the trick!
> > > >> >
> > > >> > I think I'll need to look more closely at (1) and (2) with
> Tatiana.
> > > >> >
> > > >> > Thanks,
> > > >> > John
> > > >> >
> > > >> >
> > > >> >
> > > >> >
> > > >>
> > > >>
> > > >> --
> > > >> Tracey Dorian
> > > >> NOAA/NCEP/EMC
> > > >> IMSG contractor
> > > >> 5830 University Research Court
> > > >> College Park, MD 20740
> > > >>
> > > >>
> > > >
> > >
> > >
> >
> >
> > --
> > Tracey Dorian
> > NOAA/NCEP/EMC
> > IMSG contractor
> > 5830 University Research Court
> > College Park, MD 20740
> >
> >
>
>
--
Tracey Dorian
NOAA/NCEP/EMC
IMSG contractor
5830 University Research Court
College Park, MD 20740
------------------------------------------------
More information about the Met_help
mailing list