René Zeidler created NIFI-16195:
-----------------------------------
Summary: ExtractEmailAttachments: Write part headers into
attributes
Key: NIFI-16195
URL: https://issues.apache.org/jira/browse/NIFI-16195
Project: Apache NiFi
Issue Type: Improvement
Components: Extensions
Affects Versions: 2.11.0
Environment: N/A
Reporter: René Zeidler
The {{ExtractEmailAttachments}} processor creates a FlowFile for each "part" of
an email. For an HTML email, this can be many different parts:
* The HTML body
* An alternative plain text representation of the body
* Inline images
* Actual attachments
Currently there is no easy way to filter out the parts you actually want (e.g.
just the attachments but not the inline images) except guessing from the
content of the attachment. Each part has a Content-Type header, and optional
Content-ID, Content-Description, Content-Disposition, and
Content-Transfer-Encoding headers.
This is the structure of an example email:
* Content-Type: multipart/mixed; boundary="...1..."
** Content-Type: multipart/related; type="multipart/alternative";
boundary="...2..."
*** Content-Type: multipart/alternative; boundary="...3..."
**** Content-Type: text/plain; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
_<1: plaintext body>_
**** Content-Type: text/html; charset="iso-8859-1"
Content-Transfer-Encoding: quoted-printable
_<2: html body>_
*** Content-ID: <0011223344556677@8899AABBCCDDEEFF>
Content-Type: image/png; name="inline_image.png"
Content-Disposition: inline; filename="inline_image.png"
Content-Transfer-Encoding: base64
_<3: inline image>_
** Content-Type: application/pdf; name="my-attachment.pdf"
Content-Description: my-attachment.pdf
Content-Disposition: attachment;
filename="my-attachment.pdf"; size=71521;
creation-date="Thu, 13 Aug 2026 11:02:50 GMT";
modification-date="Thu, 13 Aug 2026 11:01:24 GMT"
Content-Transfer-Encoding: base64
_<4: pdf attachment>_
The processor correctly returns all four leaf parts of the message, but without
a way to distinguish them. The simplest solution would be to write all
{{Content-*}} headers into attributes.
--
This message was sent by Atlassian Jira
(v8.20.10#820010)
