Optional decimal field coding

fplpo_x · November 15, 2006, 2:28pm

Imported from previous forum

fplpo_x · November 15, 2006, 2:28pm

[ original email was from Dmitriy Ovdienko - odp@btobits.com ]
Hi All

We have optional decimal field, defiend by the following template rule:

According $10.6.2 of the Spec v1.1 “The mantissa is present in the stream iff the exponent is not NULL”.

Should we send NULL exponent without mantissa to transfer NULL-Decimal value (pMap(exponent) = 1, pMap(mantissa) = 0)?
Should we apply operator for mantissa if exponent is NULL? If yes, the next value for mantissa will be ‘11’. Otherwise, the next value or mantissa will be ‘10’.

Regards,
Dmitriy

davidrosenborg30_x · November 15, 2006, 2:40pm

You should send NULL, pMap(exponent) = 1, and no bit at all for the mantissa

1.a: If the previous exponent was NULL, then you should send pMap(exponent) = 0, and no bit for the mantissa

No, if the exponent is NULL the mantissa operator is not used at all

/David

Hi All

We have optional decimal field, defiend by the following template rule:

According $10.6.2 of the Spec v1.1 “The mantissa is present in the
stream iff the exponent is not NULL”.

Should we send NULL exponent without mantissa to transfer NULL-
Decimal value (pMap(exponent) = 1, pMap(mantissa) = 0)?

Should we apply operator for mantissa if exponent is NULL? If yes,
the next value for mantissa will be ‘11’. Otherwise, the next value
or mantissa will be ‘10’.

Regards, Dmitriy

fplpo_x · November 15, 2006, 3:04pm

[ original email was from Dmitriy Ovdienko - odp@btobits.com ]
> 1. You should send NULL, pMap(exponent) = 1, and no bit at all for

the mantissa

So, does the size of presence map depend on transfered value? e.g the size of pmap will be increased by 1 if we should sent non-null value (a bit for exponenta and a bit for mantissa)?

rolfandersson58_x · November 15, 2006, 3:25pm

You should send NULL, pMap(exponent) = 1,
and no bit at all for the mantissa

So, does the size of presence map depend on transfered value?
e.g the size of pmap will be increased by 1 if we should sent
non-null value (a bit for exponenta and a bit for mantissa)?

Yes, that’s the intent

larsmattsson_x · November 15, 2006, 4:14pm

You should send NULL, pMap(exponent) = 1, and no bit at all for
the mantissa

So, does the size of presence map depend on transfered value?
e.g the size of pmap will be increased by 1 if we should sent non-null
value (a bit for exponenta and a bit for mantissa)?

Yes, that’s the intent

What? You mean that a pm for template 1 can be of diffrent size depening on what is in the pm?

The Fast Specification 1.x.05 says that the mantissa appears in the stream iff the exponent appears in the stream but it dosn’t say that it then also takes a spot in the PresentMap dose it?

rolfandersson58_x · November 15, 2006, 5:12pm

<>> 1. You should send NULL, pMap(exponent) = 1,

and no bit at all for the mantissa

So, does the size of presence map depend on transfered value?
e.g the size of pmap will be increased by 1 if we should sent
non-null value (a bit for exponenta and a bit for mantissa)?

Yes, that’s the intent

What? You mean that a pm for template 1 can be of diffrent size
depening on what is in the pm?

The Fast Specification 1.x.05 says that the mantissa appears in
the stream iff the exponent appears in the stream but it dosn’t
say that it then also takes a spot in the PresentMap dose it?

The pmap entry for the mantissa is not needed if the number is
null (indicated by a null exponent)

Otherwise, there is a pmap entry for the mantissa if there is an
operator specificed for the mantissa.

/Rolf

davidrosenborg30_x · November 15, 2006, 7:53pm

The spec isn’t clear enough on this. I’ve an entry on my todo list to add a note about that the absence of the mantissa includes any related slots in the presence map.

/David

You should send NULL, pMap(exponent) = 1, and no bit at all for
the mantissa

So, does the size of presence map depend on transfered value?
e.g the size of pmap will be increased by 1 if we should sent non-
null value (a bit for exponenta and a bit for mantissa)?

Yes, that’s the intent

What? You mean that a pm for template 1 can be of diffrent size depening
on what is in the pm?

The Fast Specification 1.x.05 says that the mantissa appears in the
stream iff the exponent appears in the stream but it dosn’t say that it
then also takes a spot in the PresentMap dose it?

rolfandersson58_x · November 15, 2006, 7:55pm

All subsequent slots in the same pmap …

The spec isn’t clear enough on this. I’ve an entry on my todo list to
add a note about that the absence of the mantissa includes any related
slots in the presence map.

/David

You should send NULL, pMap(exponent) = 1, and no bit at all
for the mantissa

So, does the size of presence map depend on transfered value?
e.g the size of pmap will be increased by 1 if we should sent non-
null value (a bit for exponenta and a bit for mantissa)?

Yes, that’s the intent

What? You mean that a pm for template 1 can be of diffrent size
depening on what is in the pm?

The Fast Specification 1.x.05 says that the mantissa appears in the
stream iff the exponent appears in the stream but it dosn’t say that
it then also takes a spot in the PresentMap dose it?

norbertklose_x · May 10, 2007, 12:14pm

So, if the mantissa operator is not used at all in the case when the exponent is NULL, is the previuos value for the absent mantissa set to NULL or left unchanged ?

You should send NULL, pMap(exponent) = 1, and no bit at all for
the mantissa

1.a: If the previous exponent was NULL, then you should send
pMap(exponent) = 0, and no bit for the mantissa

No, if the exponent is NULL the mantissa operator is not used at all

/David

Hi All

We have optional decimal field, defiend by the following template
rule:

According $10.6.2 of the Spec v1.1 “The mantissa is present in the
stream iff the exponent is not NULL”.

Should we send NULL exponent without mantissa to transfer NULL-
Decimal value (pMap(exponent) = 1, pMap(mantissa) = 0)?

Should we apply operator for mantissa if exponent is NULL? If yes,
the next value for mantissa will be ‘11’. Otherwise, the next value
or mantissa will be ‘10’.

Regards, Dmitriy

fplpo_x · May 15, 2007, 11:58pm

[ original email was from Scott McCaskill - smccaskill@zonetrading.com ]
My interpretation is that the previous value for the mantissa is unchanged in that case. Only the exponent is nullable; if the exponent is encoded as NULL, then the mantissa is not encoded at all (not even as the null value), so I wouldn’t expect its dictionary value to change.

So, if the mantissa operator is not used at all in the case when the
exponent is NULL, is the previuos value for the absent mantissa set to
NULL or left unchanged ?

You should send NULL, pMap(exponent) = 1, and no bit at all for the
mantissa

1.a: If the previous exponent was NULL, then you should send
pMap(exponent) = 0, and no bit for the mantissa

No, if the exponent is NULL the mantissa operator is not used
at all

/David

Hi All

We have optional decimal field, defiend by the following template
rule:

According $10.6.2 of the Spec v1.1 “The mantissa is present in the
stream iff the exponent is not NULL”.

Should we send NULL exponent without mantissa to transfer NULL-
Decimal value (pMap(exponent) = 1, pMap(mantissa) = 0)?

Should we apply operator for mantissa if exponent is NULL? If
yes, the next value for mantissa will be ‘11’. Otherwise, the
next value or mantissa will be ‘10’.

Regards, Dmitriy

davidrosenborg30_x · May 16, 2007, 7:31am

This is correct. The operator of the mantissa is not exercised at all in the case where the exponent is NULL. As a consequence, the previous value is left untouched.

My interpretation is that the previous value for the mantissa is
unchanged in that case. Only the exponent is nullable; if the exponent
is encoded as NULL, then the mantissa is not encoded at all (not even as
the null value), so I wouldn’t expect its dictionary value to change.

So, if the mantissa operator is not used at all in the case when the
exponent is NULL, is the previuos value for the absent mantissa set to
NULL or left unchanged ?

Topic		Replies	Views
Optional decimal fields with individual operators imported , fast-protocol , xid_169680	13	34	September 1, 2008
Decimal Encoding and Presence Map imported , fast-protocol , xid_169539	5	37	November 9, 2007
Decimal Presence Map Utilization imported , fast-protocol , xid_169585	17	44	August 28, 2007
Decimal Field Instruction having single field operator imported , fast-protocol , xid_169766	6	99	September 11, 2009
Optional Decimal Logic imported , fast-protocol , xid_169759	4	45	June 30, 2009

Optional decimal field coding

Related topics