FT-Index und PDF

[ThomasF]

Hallo,

ich bekomme angehängte PDF's nicht in den FT-Index. Wenn ich die als OLE und "Anzeigen als" Rich Text einstelle, dann findet der Indexer das Dokument (aber auch NUR dann).
Das Häkchen beim Index "Angehängte Dateien indizieren" ist gesetzt und die Option der Konvertierungsfilter scheint auch egal.

[datenbanken24]

Hallo Thomas,

wir kennen das "Problem" auch.
Und das muß nicht mal an Notes liegen, sondern an den PDFs selbst.

Wir haben mehrere Datenbanken mit Datei-Indizierung,
die PDFs enthalten. Manche PDFs in der gleichen Datenbank werden textlich über den FT erkannt - und manche nicht.

Im einfachsten Fall bestehen die PDFs gar nicht aus Text sondern aus eingescantem Text und damit aus einem Bild, wodurch eine FT-Indizierung unmöglich ist.

Aber auch bei hundertprozentig textlichen PDFs, werden einige PDFs nicht mit indiziert. Wir sind sicher, dass das an den PDFs selbst liegt, wissen aber noch nicht woran genau.

Es wird sowas sein, wie das man einige JPG-Bilder auch pardu nicht im Browser angezeigt bekommt. Das ist dann auch eine Einstellung bzw. eine Sonderform im JPG und man kann lange suchen, warum denn der Browser dieses eine Bild nicht anzeigt.

Gruß,
Uwe

[datenbanken24]

PS:
Die conversion filters option setzen wir immer.
Ohne die (Default) klappt in Hinsicht Dateianhänge fast gar nix.

[ThomasF]

Hab' ich mir fast gesdacht.
Ok danke für eure Mühe.

[Glombi]

siehe dazu:

Adobe .PDF files are not full text indexed as expected
Product:
Lotus Domino  >  Lotus Domino Server  >  Versions 6.5, 6.0
Platform(s):
Platform Independent
Doc Number:
1089010
Published:  26.08.2005
Technote

Problem

It has been found that Adobe .PDF files that have security settings are not being full text indexed as expected, even though the files are version 1.2 or above.  Specifically, a full text index of version 1.2 PDF files does not complete successfully unless, under security, everything is allowed.  On version 1.3 PDF files, they will be full text indexed only if you allow "Can Highlight Text" on the .PDF file security.



Solution
This issue was reported to Quality Engineering as SPR# FBEE4SVURB and has been addressed in Domino 6.0.5/6.5.4.

Excerpt from the Lotus Notes and Domino Release 6.5.4 and 6.0.5 MR fix list (available at http://www.ibm.com/developerworks/lotus/):

Full-text indexing
SPR# FBEE4SVURB - Adobe .pdf files were not being full text indexed. This problem has been fixed in 6.0.5 and 6.5.4. This regression was introduced in 6.0. 

Supporting Information:
Indexing for Adobe .PDF files is handled by a third-party filter.  It appears that both .PDF versions 1.2 and 1.3 will work, as long as security options are set to "Allow selecting text and graphics" AND there is no security password setting.

To check this in Acrobat, go to File -> Document Properties -> Description.  The relevant information is in the PDF Version field.

Adobe PDF 1.4 support
It was originally documented that Notes/Domino 6.x supported Adobe PDF 1.4.  However, research by IBM Lotus Quality Engineering and Verity later found that the filter did not work as expected for all the content variations within the file type and, as a result, it is no longer documented as a supported file type.  There may be variations of the file type that will work as expected, but this does not mean that all variations will work.  Also the Adobe Acrobat release used to save the 1.4 PDF file, as well as the Notes release used to index the attachment, can also result in variations in behavior.  For additional information on this topic, refer to the document titled "Not All PDF Version 1.x Documents Are Supported for Full Text Indexing" (#1188610).

[Glombi]

Are PDF version 1.5 (Adobe 6) documents supported for full text indexing?
Product:
Lotus Domino  >  Lotus Domino Administrator  >  Versions 7.0, 6.5, 6.0, 5.0
Platform(s):
Platform Independent
Doc Number:
1199681
Published:  05.11.2005
Technote

Problem

You are using Adobe 6 and would like support for searching documents created in the Adobe 6 PDF 1.5 format. Is this format supported?



Solution
Currently the Verity search filters support only up to PDF 1.3. The 1.5 format is not supported.

This issue has been reported to Quality Engineering as SPR# WPAS69WC65 as an enhancement request to support searching on the 1.5 format.

Saving the files in Adobe 6, using the PDF 1.3 format, is currently the only workaround.

Supporting Information:
A separate issue that involved full text indexing failing on supported PDF formats occurred in cases where the file security settings had certain options enabled. For details on this issue, including fix release information, refer to the document titled "Adobe PDF files are not full text indexed as expected" (#1089010).

Adobe PDF 1.4 (and later) support:
It was originally documented that Notes/Domino 6.x supported Adobe PDF 1.4. However, research by IBM Lotus Quality Engineering and Verity later found that the filter did not work as expected for all the content variations within the file type and, as a result, it is no longer documented as a supported file type. With PDF 1.4 (and later versions) there may be variations of the file type that will work as expected, but this does not mean that all variations will work. Also, the Adobe Acrobat release used to save the PDF file, as well as the Notes release used to index the attachment, can also result in variations in behavior. For additional information on this topic, refer to the document titled "Not all PDF version 1.x documents are supported for Full Text Indexing" (#1089010).

[datenbanken24]

Danke für die Info.

Da wird natürlich nun einiges klar.

Leider seeehr ernüchternde Nachrichten,

"...supported erst ab 1.4 aber nicht mehr für 1.5 ..."

Wie sagt man sowas nun wieder seinen Kunden... 

[mastertom]

Mir ist ein Gerücht von einem Bekannten berichtet worden, dass besagt, dass das Installieren von Adobe auf dem Domino-Server das Problem beheben soll.

Angeblich wird so die Notes-Indizierung durch den Adobe Acrobat ersetzt?

Hat das schon mal jemand ausprobiert? Klingt nicht realistisch, oder doch?