All,
I need to store C++ containers like maps , vectors into hdf5 . One method is to use variable length datatypes . To my knowledge too much use of variable length datatype effects the performance . Is there a better way to do this ?
Regards
Ramakrishnan
Hi Quincey ,
Consider a map ( key,value pair) having three elements . In order to
store this I can create a compound data type containing key and value
types in it . Then I need to create a two dimensional dataset , wherein
each row will give me a map entry . Since I said three elements, I can
make the dataspace dimension as three. This will be equivalent to a 2x3
matrix .
Is my understanding right?
In my case what I need to store together is a map and a pointer (
dataset region reference ) . I need to store only one reference along
with entire map contents . How can I achieve this . Any suggestions ?
Regards,
Ramakrishnan
···
-----Original Message-----
From: Quincey Koziol [mailto:koziol@hdfgroup.org]
Sent: Wednesday, April 30, 2008 12:01 AM
To: Ramakrishnan Iyer
Cc: hdf-forum@hdfgroup.org; HDF Helpdesk
Subject: Re: [hdf-forum] Best way of representing C++ containers in hdf5
Hi Ramakrishnan,
On Apr 29, 2008, at 7:31 AM, Ramakrishnan Iyer wrote:
All,
I need to store C++ containers like maps , vectors into hdf5 . One
method is to use variable length datatypes . To my knowledge too
much use of variable length datatype effects the performance . Is
there a better way to do this ?
Most (all?) C++ containers allow you to determine the number of
objects within them. You could always use that count to define a
dataspace dimension and then store the objects as elements of the
dataset...
Quincey
----------------------------------------------------------------------
This mailing list is for HDF software users discussion.
To subscribe to this list, send a message to
hdf-forum-subscribe@hdfgroup.org.
To unsubscribe, send a message to hdf-forum-unsubscribe@hdfgroup.org.
Hi Ramakrishnan,
Hi Quincey ,
Consider a map ( key,value pair) having three elements . In order to
store this I can create a compound data type containing key and value
types in it . Then I need to create a two dimensional dataset , wherein
each row will give me a map entry . Since I said three elements, I can
make the dataspace dimension as three. This will be equivalent to a 2x3
matrix .
Is my understanding right?
If you create a compound datatype that holds both the key & value, why do you need a two dimensional dataset?
In my case what I need to store together is a map and a pointer (
dataset region reference ) . I need to store only one reference along
with entire map contents . How can I achieve this . Any suggestions ?
Hmm, what about putting the reference as an attribute on the dataset that contains the contents of the map?
Quincey
···
On Apr 30, 2008, at 1:45 AM, Ramakrishnan Iyer wrote:
Regards,
Ramakrishnan
-----Original Message-----
From: Quincey Koziol [mailto:koziol@hdfgroup.org]
Sent: Wednesday, April 30, 2008 12:01 AM
To: Ramakrishnan Iyer
Cc: hdf-forum@hdfgroup.org; HDF Helpdesk
Subject: Re: [hdf-forum] Best way of representing C++ containers in hdf5
Hi Ramakrishnan,
On Apr 29, 2008, at 7:31 AM, Ramakrishnan Iyer wrote:
All,
I need to store C++ containers like maps , vectors into hdf5 . One
method is to use variable length datatypes . To my knowledge too
much use of variable length datatype effects the performance . Is
there a better way to do this ?
Most (all?) C++ containers allow you to determine the number of
objects within them. You could always use that count to define a
dataspace dimension and then store the objects as elements of the
dataset...
Quincey
----------------------------------------------------------------------
This mailing list is for HDF software users discussion.
To subscribe to this list, send a message to
hdf-forum-subscribe@hdfgroup.org.
To unsubscribe, send a message to hdf-forum-unsubscribe@hdfgroup.org.
----------------------------------------------------------------------
This mailing list is for HDF software users discussion.
To subscribe to this list, send a message to hdf-forum-subscribe@hdfgroup.org.
To unsubscribe, send a message to hdf-forum-unsubscribe@hdfgroup.org.