Re: [ferret_users] defining time axis

["Ansley C. Manke" <ansley.b.manke@xxxxxxxx>]

Hi Arijeet,

To define a mask that picks out only the April months from each year you 
would want this, wouldn't you?

  let mask_april = if yr_months EQ 4 then 1

Then before applying the mask, replace the missing-value flags in your 
data with some other known value which will never appear in the valid data.

  let u_adcp_2 = MISSING(u_adcp, -1.e10)

  let u_adcp_april = u_adcp_2* mask_april
  let u_adcp_compr = COMPRESSL(u_adcp_april )

The variable u_adcp_april has the default Ferret missing-value flag so 
compressing it uses that flag, and will keep all other points including 
any values of -1.e10.

   define axis/t=1:11:1 t11
   let/bad=-1.e10 u_adcp_april_all = u_adcp_compr[gt=t11@ASN]

and this should be all the april data, with any months that originally 
had missing data, still containing a missing-value flag.

(I'm just typing commands here, so of course there may be typo's!)

-Ansley

On 1/4/2017 11:42 PM, Arijeet Dutta wrote:
> Hi all
> I have monthly current data (observation) from 2001 to 2011, with many
> of the months of 2009 having no observation and hence no data. I
> wanted to produce a plot where I can get only the April months from
> 2001 to 2011 in the time axis. So I found posts helpful from ferret
> archive and I did the following
>
> yes? yes?      let L_indx    = L[GT=u_adcp]
> yes? yes?      let yr_months = L_indx- 12*INT((L_indx-1)/12)
> yes? yes?      let mask_other_months = IF yr_months GT 3 AND yr_months
> LT 5 THEN 1
> yes? yes?      let u_adcp_masked = u_adcp * mask_other_months
> yes? yes?      let u_adcp_compr  = COMPRESSL(u_adcp_masked)
>
> Now u_adcp_compr gives 10 data points since 2009 data for april is
> missing. But I want a plot where 2009 april data should show missing
> value but when I am doing
>
> yes? yes?      define axis/t=1:11:1 t11
> yes? yes?      let u_adcp_april_all = u_adcp_compr[gt=t11@ASN]
>
> u_adcp_april_all shows 11th data points with the missing one being the
> 11th one but I want 9th one to be missing; how can I better define the
> time axis to avoid the problem??
> Thanks
> Arijeet