philipp/hdf5
1
0
Fork 0
Code Issues Pull Requests Packages Projects Releases Wiki Activity

[svn-r3659]

Browse Source 
Added html files with references to the fortran 90 files.
(under examples/)
This commit is contained in:
Barbara Jones
2001-03-19 09:26:47 -05:00
parent 380eb42bc5
commit f11cef2c56
30 changed files with 2967 additions and 3139 deletions
Download Patch File Download Diff File Expand all files Collapse all files

346
doc/html/Tutor/mount.html
View File

@@ -35,244 +35,196 @@ width=78 height=27 alt="NCSA"><P></A>
<A NAME="def">
<H2>Mounting Files</H2>
HDF5 allows you to combine two or more HDF5 files in a manner similar to mounting
files in UNIX. The group structure and metadata from one file appear as though
HDF5 allows you to combine two or more HDF5 files in memory
in a manner similar to mounting files in UNIX.
The group structure and metadata from one file appear as though
they exist in another file. The following steps are involved:
<OL>
<LI>Open the files.
<LI>Choose the "mount point" in the first file (parent). The "mount point" in
HDF5 is a group ( it can also be the root group ).
<LI>Choose the <strong>mount point</strong> in the first file
(the parent file). The mount point in
HDF5 is a group, which CANNOT be the root group.
<LI>Use the HDF5 API function H5Fmount to mount the second file (child) in
the first one.
<LI>Use the HDF5 routine <CODE>H5Fmount</CODE> / <CODE>h5fmount_f</CODE>
to mount the second file (the child file) in the first file.
<LI>Work with the objects in the second file as if they were members of
the "mount point" group in the first file. The previous contents of
the "mount point" group are temporarily hidden.
the mount point group in the first file. The previous contents of
the mount point group are temporarily hidden.
<LI>Unmount the second file using the H5Funmount function when the work is done.
<LI>Unmount the second file using <CODE>H5Funmount</CODE> /
<CODE>h5funmount_f</CODE> when the work is done.
</OL>
<H2> Programming Example</H2>
<A NAME="desc">
<H3><U>Description</U></H3>
In the following example we create one file with a group in it, and another
file with a dataset. Mounting is used to access the dataset from the second
file as a member of a group in the first file. The following picture
illustrates this concept.
In the following example, we create one file containing a group and
another file containing a dataset.
Mounting is used to access the dataset from the second
file as a member of a group in the first file.
The following figures illustrate this concept.
<PRE>
'FILE1' 'FILE2'
FILE1 FILE2
-------------------- --------------------
! ! ! !
! / ! ! / !
! | ! ! | !
! | ! ! | !
! V ! ! V !
! -------- ! ! ---------- !
! ! Group ! ! ! ! Dataset! !
! --------- ! ! ---------- !
!------------------! !------------------!
-------------------- --------------------
! ! ! !
! / ! ! / !
! | ! ! | !
! | ! ! | !
! V ! ! V !
! -------- ! ! ---------- !
! ! Group ! ! ! ! Dataset! !
! --------- ! ! ---------- !
!------------------! !------------------!
</PRE>
After mounting the second file, 'FILE2', under the group in the file, 'FILE1',
the parent has the following structure:
After mounting <code>FILE2</code> under the group in <code>FILE1</code>,
the parent file has the following structure:
<PRE>
'FILE1'
FILE1
--------------------
! !
! / !
! | !
! | !
! V !
! -------- !
! ! Group ! !
! --------- !
! | !
! | !
! V !
! ----------- !
! ! Dataset ! !
! !---------- !
! !
!------------------!
--------------------
! !
! / !
! | !
! | !
! V !
! -------- !
! ! Group ! !
! --------- !
! | !
! | !
! V !
! ----------- !
! ! Dataset ! !
! !---------- !
! !
!------------------!
</PRE>
[ <A HREF="examples/h5_mount.c">Download h5_mount.c</A> ]
<PRE>
+++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
[ <A HREF="examples/h5_mount.c">C program</A> ]
- <code>h5_mount.c</code><BR>
[ <A HREF="examples/mount.f90">FORTRAN program</A> ]
- <code>mountexample.f90</code>
<P>
#include&lt;hdf5.h&gt;
<B>NOTE:</B> To download a tar file of the examples, including a Makefile,
please go to the <A HREF="references.html">References</A> page.
#define FILE1 "mount1.h5"
#define FILE2 "mount2.h5"
#define RANK 2
#define NX 4
#define NY 5
int main(void)
{
hid_t fid1, fid2, gid; /* Files and group identifiers */
hid_t did, tid, sid; /* Dataset and datatype identifiers */
herr_t status;
hsize_t dims[] = {NX,NY}; /* Dataset dimensions */
int i, j;
int bm[NX][NY], bm_out[NX][NY]; /* Data buffers */
/*
* Initialization of buffer matrix "bm"
*/
for(i =0; i &lt; NX; i++) {
for(j = 0; j &lt; NY; j++)
bm[i][j] = i + j;
}
/*
* Create first file and a group in it.
*/
fid1 = H5Fcreate(FILE1, H5F_ACC_TRUNC, H5P_DEFAULT, H5P_DEFAULT);
gid = H5Gcreate(fid1, "/G", 0);
/*
* Close group and file
*/
H5Gclose(gid);
H5Fclose(fid1);
/*
* Create second file and dataset "D" in it.
*/
fid2 = H5Fcreate(FILE2, H5F_ACC_TRUNC, H5P_DEFAULT, H5P_DEFAULT);
dims[0] = NX;
dims[1] = NY;
sid = H5Screate_simple(RANK, dims, NULL);
did = H5Dcreate(fid2, "D", H5T_NATIVE_INT, sid, H5P_DEFAULT);
/*
* Write data to the dataset.
*/
status = H5Dwrite(did, H5T_NATIVE_INT, H5S_ALL, H5S_ALL, H5P_DEFAULT, bm);
/*
* Close all identifiers.
*/
H5Sclose(sid);
H5Dclose(did);
H5Fclose(fid2);
/*
* Reopen both files
*/
fid1 = H5Fopen(FILE1, H5F_ACC_RDONLY, H5P_DEFAULT);
fid2 = H5Fopen(FILE2, H5F_ACC_RDONLY, H5P_DEFAULT);
/*
* Mount second file under G in the first file.
*/
H5Fmount(fid1, "/G", fid2, H5P_DEFAULT);
/*
* Access dataset D in the first file under /G/D name.
*/
did = H5Dopen(fid1,"/G/D");
tid = H5Dget_type(did);
status = H5Dread(did, tid, H5S_ALL, H5S_ALL, H5P_DEFAULT, bm_out);
/*
* Print out the data.
*/
for(i=0; i &lt; NX; i++){
for(j=0; j &lt; NY; j++)
printf(" %d", bm_out[i][j]);
printf("\n");
}
/*
* Close all identifers
*/
H5Tclose(tid);
H5Dclose(did);
/*
* Unmounting second file
*/
H5Funmount(fid1, "/G");
/*
* Close both files
*/
H5Fclose(fid1);
H5Fclose(fid2);
return 0;
}
++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++++
</PRE>
<A NAME="rem">
<H3><U>Remarks</U></H3>
<UL>
<LI> The first part of the program creates a group in one file and creates
and writes a dataset to another file.
<P>
<LI> Both files are reopened with the H5F_ACC_RDONLY access flag since no
objects will be modified. The child should be opened with H5F_ACC_RDWR if
the dataset is to be modified.
<LI> Both files are reopened and the second file is mounted in the first
using <CODE>H5Fmount</CODE> / <CODE>h5fmount_f</CODE>.
If no objects will be modified, the
files can be opened with <CODE>H5F_ACC_RDONLY</CODE>
(<CODE>H5F_ACC_RDONLY_F</CODE> in FORTRAN).
If the data is to be modified, the files should be opened with
<CODE>H5F_ACC_RDWR</CODE> (<CODE>H5F_ACC_RDWR_F</CODE> in FORTRAN).
<P>
<LI> The second file is mounted in the first using the H5Fmount function.
<PRE>
herr_t H5Fmount(hid_t loc_id, const char *name, hid_t file_id, hid_t plist_id)
</PRE>
<UL>
<LI> The first two arguments specify the location of the "mount point"
( a group ). In this example the "mount point" is a group "/G" in the
file specified by its handle <I>fid1</I>. Since group G is in the root
group of the first file, one can also use just "G" to identify it.
<I><B>C:</B></I>
<pre>
herr_t H5Fmount (hid_t loc_id, const char *dsetname,
hid_t file_id, hid_t access_prp)
</pre>
<P>
Below is a description of another scenario:<BR>
Suppose group G was a member of group D in the first file (<I>fid1</I>).
Then mounting point G can be specified in two different ways:
<I><B>FORTRAN:</B></I>
<pre>
h5fmount_f (loc_id, dsetname, file_id, hdferr, access_prp)
loc_id IN: INTEGER (HID_T)
dsetname IN: CHARACTER (LEN=*)
file_id IN: INTEGER (HID_T)
hdferr OUT: INTEGER
access_prp IN: INTEGER (HID_T), OPTIONAL
(Default value: H5P_DEFAULT_F)
</pre>
<P>
<UL>
<LI> <I>loc_id</I> is <I>fid1</I><BR>
<I>name</I> is "D/G"
<LI> The <em>loc_id</em> and <em>dsetname</em> arguments
specify the location of the mount point.
In this example, the mount point is a group <code>/G</code> in the
specified file. Since the group <code>/G</code> is in the root
group of the first file, one can also use just <code>G</code> to
identify it.
<P>
<LI> <I>loc_id</I> is an identifier of the group "D"<BR>
<I>name</I> is just "G"
Below is a description of another scenario:
<p>
Suppose the group <code>G</code> were a member of
the group <code>H</code> in the first file.
Then the mount point <code>G</code> can be specified in
two different ways:
<P>
<UL>
<LI> <em>loc_id</em> is the file identifier for the first file.<BR>
<em>dsetname</em> is <code>H/G</code>.
<P>
<LI> <em>loc_id</em> is the identifier for the group <code>H</code>.<BR>
<em>dsetname</em> is <code>G</code>.
</UL>
<P>
<LI> The third argument is an identifier of the file which will be mounted.
Only one file can be mounted per "mount point".
<LI> The <em>file_id</em> argument is the identifier for the file
which will be mounted.
Only one file can be mounted per mount point.
<P>
<LI> The fourth argument is an identifier of the property list to be used.
Currently, only the default property list, H5P_DEFAULT, can be used.
<LI> The <I>access_prp</I> argument is the identifier for the property list
to be used. Currently, only the default property list,
<CODE>H5P_DEFAULT</CODE>, can be used in C.
In FORTRAN, this argument can be omitted or
<CODE>H5P_DEFAULT_F</CODE> can be used.
<P>
<LI> The C function <CODE>H5Fmount</CODE> returns a non-negative
value if successful and a negative value otherwise.
With the FORTRAN routine, <CODE>h5fmount_f</CODE>,
the return value of the call is returned in <em>hdferr</em>:
0 if successful and -1 otherwise.
</UL>
<P>
<LI> In this example we just read data from the dataset D. One can modify
data also. If the dataset is modified while the file is mounted, it becomes
modified in the original file too after the file is unmounted.
<LI>In this example, we only read data from the dataset <code>D</code>.
One can also modify data.
If the dataset is modified while the file is mounted, it is
modified in the original file after the file is unmounted.
<P>
<LI> The file is unmounted with the H5Funmount function:
<PRE>
herr_t H5Funmount(hid_t loc_id, const char *name)
</PRE>
Arguments to this function specify the location of the "mount point".
In our example <I>loc_id</I> is an identifier of the first file, and
<I>name</I> is the name of group G, "/G". One could also use "G"
instead of "/G" since G is a member of the root group in the file
<I>fid1</I>. Notice that H5Funmount does not close files. They are closed
with the respective calls to the H5Fclose function. Closing the parent
automatically unmounts the child.
<LI> The file is unmounted with <CODE>H5Funmount</CODE> /
<CODE>h5funmount_f</CODE>:
<P>
<I><B>C:</B></I>
<pre>
herr_t H5Funmount (hid_t loc_id, const char *dsetname)
</pre>
<P>
<I><B>FORTRAN:</B></I>
<pre>
h5funmount_f (loc_id, dsetname, hdferr)
loc_id IN: INTEGER (HID_T)
dsetname IN: CHARACTER (LEN=*)
hdferr OUT: INTEGER
</pre>
<P>
<ul>
<li>The <I>loc_id</I> and <I>dsetname</I> arguments specify the location
of the mount point.
In our example <I>loc_id</I> is the first file's file identifier
and <I>dsetname</I> is the name of group <code>/G</code>.
</ul>
<P>
<LI>The h5dump utility cannot display files in memory, therefore no output
of FILE1 after FILE2 was mounted is provided.
<li>Note that <CODE>H5Funmount</CODE> / <CODE>h5funmount_f</CODE>
does not close files. Files are closed with the respective calls to
the <CODE>H5Fclose</CODE> / <CODE>h5fclose_f</CODE> function.
<P>
<li>Closing the parent file automatically unmounts the child file.
<P>
<LI>The <code>h5dump</code> utility cannot display files in memory.
Therefore, no output of <code>FILE1</code> after <code>FILE2</code>
was mounted is provided.
</UL>
</UL>
@@ -291,8 +243,10 @@ int main(void)
<!-- <A HREF="helpdesk.mail.html"> -->
<A HREF="mailto:hdfhelp@ncsa.uiuc.edu">
hdfhelp@ncsa.uiuc.edu</A>
<BR> <H6>Last Modified: August 27, 1999</H6><BR>
<br>
<BR> <H6>Last Modified: March 16, 2001</H6><BR>
<!-- modified by Barbara Jones - bljones@ncsa.uiuc.edu -->
<!-- modified by Frank Baker - fbaker@ncsa.uiuc.edu -->
</FONT>
<BR>
<!-- <A HREF="mailto:hdfhelp@ncsa.uiuc.edu"> -->