[Solved] How can i get the Object Id or the Path of Blobfile?

The blobfile lives here:


How can i find this Blobfile in my Portal Catalog or in my Site Tree? Can i convert the "0x..." in an oid or what is the right way to resolve physical Path to the Object?

Now, i have a browserview to check my candidates:

class Clean(BrowserView):

    def __call__(self):
        from ZODB.utils import p64
        from ZODB.blob import FilesystemHelper
        portal = api.portal.get()
        app = portal.aq_parent
        fshelper = FilesystemHelper("/Plone5/py2/var/blobstorage")
        mylist = [
        for oid, oidpath in fshelper.listOIDs():
            if oidpath in mylist:
                obj = app._p_jar.get(oid)
                # obj : <ZODB.blob.Blob object at 0x7fa4afd73350 oid 0xa7c4b4 in <Connection at 7fa4b4e0fc10>>
                # how to continue ? 
        return 'ok'
1 Like

You cannot easily.

The hex values specify the oid and the serial (first and second 8 bytes) of the object. With the oid, you can load the object from the ZODB (as you did). But, this gives you the raw object -- without its context, e.g. its location.

For a client, I have implemented a small script which visits all File objects and outputs a map "location -> blob-oid". A similar script would allow you to locate your blobs based on their oid.

Something like this:

for brain in context.portal_catalog(portal_type='File'):
    obj = brain.getObject()
    path_on_disk = obj.file._blob._p_blob_committed
    url = obj.absolute_url()

I'm unware of something that will do the reverse as the zodb object doesn't have __parent__ pointers:

>>> obj = app._p_jar.get(b'\x00\x00\x00\x00\x00\x00\x1a6')
>>> obj
<ZODB.blob.Blob object at 0x11575c5d0 oid 0x1a36 in <Connection at 114bb1a10>>
>>> obj.__parent__
Traceback (most recent call last):
  File "<console>", line 1, in <module>
AttributeError: 'Blob' object has no attribute '__parent__'

Yes. I did not yet know _p_blob_committed, but if it gives your the path, it is perfect.

strange behavior :open_mouth: is there lazy load like mechanism ?

brains = catalog(portal_type=['File'])
obj = brains[0].getObject()        
blobfile = obj.file
blob = blobfile._blob
import pdb

# first attempt
(Pdb++) pp blob._p_blob_committed 

(Pdb++) dir(blob)
['__class__', '__delattr__', '__dict__', '__doc__', '__format__', '__getattribute__', '__getstate__', '__hash__', '__implemented__', '__init__', '__module__', '__new__', '__providedBy__', '__provides__', '__reduce__', '__reduce_ex__', '__repr__', '__setattr__', '__setstate__', '__sizeof__', '__str__', '__subclasshook__', '__weakref__', '_create_uncommitted_file', '_p_activate', '_p_blob_committed', '_p_blob_ref', '_p_blob_uncommitted', '_p_changed', '_p_deactivate', '_p_delattr', '_p_estimated_size', '_p_getattr', '_p_invalidate', '_p_jar', '_p_mtime', '_p_oid', '_p_serial', '_p_setattr', '_p_state', '_p_status', '_p_sticky', '_uncommitted', 'closed', 'committed', 'consumeFile', 'open', 'opened', 'readers', 'writers']

# second attempt
(Pdb++) pp blob._p_blob_committed 

My Solution open/close the file:

brains = catalog(portal_type=['File'])
obj = brains[0].getObject()        
blobfile = obj.file
blob = blobfile._blob
f = blob.open('r')
print "{} {}".format(blob._p_blob_committed, obj.absolute_url())

Now i have both, real filesystempath and object url :slight_smile:
thanks guys, for help!

Background: if i rebuild the catalog, very old word files breaks the wvText with segmentation fault. i have a script wich collect all broken word files. i will investigate what happen with this files. but i need the object path

# the bash script to find the broken word files
find /Plone5/py2/var/blobstorage/ -type f -exec sh -c "file -i {} | grep 'word' | sed 's/:.*//'" \; | while read I ; do
    wvText $I /tmp/wv.txt
    if [[ $RESULT -ne 0 ]]
      echo "$I" >> wvfiles.txt

The _p_blob_committed doesn't un-ghost of the _p_?
You'll have to https://zodb.readthedocs.io/en/latest/api.html#transaction.IPersistent._p_activate the object first: blob._p_activate() .

1 Like

Yes, that do the job.

brains = catalog(portal_type=['File'])
obj = brains[0].getObject()        
blobfile = obj.file
blob = blobfile._blob
print "{} {}".format(blob._p_blob_committed, obj.absolute_url())