Add cc bcc csv comma delimited V2

[holstbone]

@liamlynch-data - I created a clean, new pull request here.
(original PR here )

NOTES:

I did NOT incorporate changing to SMTP.send_message(msg), for two reasons:

1. The sendmail method was selected by the Plugin dev team, so updates should probably best be led by them.
2. There may be some ramifications to how attachments are managed.

Also, I did not incorporate hard coded CC or BCC fields, primarily because the UI/UX gets really messy. Since "Recipient" field is always pulled from the dataset, then the best UI/UX is to also always pull the "cc" and "bcc" from the dataset as well.

[liamlynch-data]

That's much better - thanks for the changes!

Re:
I did NOT incorporate changing to SMTP.send_message(msg)
Ok no worries - perhaps we'll change it later, it doesn't matter too much

Also, I did not incorporate hard coded CC or BCC fields
Not really an issue either

There are a couple more things to address here - if you can make those changes then I'll do a full test on it (I gave it a quick test already, seems fine once I fixed that double line issue).

[holstbone]

@liamlynch-data - Just made final adjustments!

[holstbone]

All changes from the sendmail_changes.patch have now been incorporated!

[liamlynch-data]

LGTM and tested. The following tests were run with the patch code applied:

Basic tests

Default: Normal email, no CC/BCC, Email body in table (text), no attachments ✅

Add CC/BCC fields ✅
With Email arrays ✅
Subject in column (e.g. status) ✅
Sender in column / value ✅

Error validation
Non-existent column for all the below: ✅
Recipient Email
Sender
CC
BCC
Email body
Subject

Values:
No body - ✅
No subject - ✅
No sender - ✅

Email body
(Email body in table)
Email body plain text - with a variable ✅
Email body html - with a variable ✅
Conditional formatting - SKIP - not relevant to change

Attachments
Excel attachment ✅
Tab delimited ✅
Comma delimited ✅

Email body plain text - with attach variable - not relevant, SKIP
Email body html - with attach variable - not relevant, SKIP

Regression
Recipe in project exported from an old version still works - with CSV tab delimited attachments ✅

Channels ✅
Repeated above via direct SMTP settings and a channel

CC/BCC ✅
Used an AWS SES mail channel and the dummy recipient address indicated in the messaging MTP AWS section as the sender to test CC and BCC emails (routed to addressed to me, an approved recipient) would get routed properly in the real world

[liamlynch-data]

Approving - see previous comment for details

[liamlynch-data]

@alexbourret @MayeulRousselet - as discussed it would be great if one of you could take a quick look, given the validation changes and the issues with multiple recipients we fixed

[alexbourret]

LGTM, just update the plugin.json with the new tags and category, and maybe clarify the expected format in the cc and bcc columns

[RoyTeddy]

Regression found:
if Subject (column) or Body (column) is a datetime column we have this error:
[18:05:25] [INFO] [dku.utils] - TypeError: unsupported operand type(s) for +: 'datetime.datetime' and 'str'
Works with previous version 1.0.3 in release Version: 14.3.2 here https://released-design.qa.managedinstances.dkucloud-dev.com/projects/TEDDY_COPY/flow/

error caused by the input dataset which contains an empty row so no bug here 😅

[RoyTeddy]

Found an issue but already there before:
in the Body with custom value, if we specify a dataset name with numbers like {{ attachments.16rows.html_table }}
-> beautiful error

[2026/01/20-18:25:13.499] [null-err-72] [INFO] [dku.utils]  - *************** Recipe code failed **************
[2026/01/20-18:25:13.500] [null-err-72] [INFO] [dku.utils]  - Begin Python stack
[2026/01/20-18:25:13.500] [null-err-72] [INFO] [dku.utils]  - Traceback (most recent call last):
[2026/01/20-18:25:13.501] [null-err-72] [INFO] [dku.utils]  -   File "/Users/teddy.roy@dataiku.com/devenv/DSS14/dss-home/jobs/LEGACYCFFOREXPORT/Build_send_review__NP__2026-01-20T17-25-12.075/compute_send_review_NP/custom-python-recipe/pyout4nbErQ4gweO6/python-exec-wrapper.py", line 204, in <module>
[2026/01/20-18:25:13.501] [null-err-72] [INFO] [dku.utils]  -     exec(f.read())
[2026/01/20-18:25:13.501] [null-err-72] [INFO] [dku.utils]  -   File "<string>", line 166, in <module>
[2026/01/20-18:25:13.502] [null-err-72] [INFO] [dku.utils]  -   File "/Users/teddy.roy@dataiku.com/devenv/DSS14/dss-home/code-envs/python/plugin_sendmail_managed/lib/python3.9/site-packages/jinja2/environment.py", line 1092, in from_string
[2026/01/20-18:25:13.502] [null-err-72] [INFO] [dku.utils]  -     return cls.from_code(self, self.compile(source), gs, None)
[2026/01/20-18:25:13.503] [null-err-72] [INFO] [dku.utils]  -   File "/Users/teddy.roy@dataiku.com/devenv/DSS14/dss-home/code-envs/python/plugin_sendmail_managed/lib/python3.9/site-packages/jinja2/environment.py", line 757, in compile
[2026/01/20-18:25:13.503] [null-err-72] [INFO] [dku.utils]  -     self.handle_exception(source=source_hint)
[2026/01/20-18:25:13.503] [null-err-72] [INFO] [dku.utils]  -   File "/Users/teddy.roy@dataiku.com/devenv/DSS14/dss-home/code-envs/python/plugin_sendmail_managed/lib/python3.9/site-packages/jinja2/environment.py", line 925, in handle_exception
[2026/01/20-18:25:13.504] [null-err-72] [INFO] [dku.utils]  -     raise rewrite_traceback_stack(source=source)
[2026/01/20-18:25:13.504] [null-err-72] [INFO] [dku.utils]  -   File "<unknown>", line 4, in template
[2026/01/20-18:25:13.504] [null-err-72] [INFO] [dku.utils]  - jinja2.exceptions.TemplateSyntaxError: expected token 'end of print statement', got 'rows'

it is a Python limitation.
Attributes that start with a number, such as "123property" have to be expressed as {{customer['123property']}}

so works with {{ attachments['16rows'].html_table }}
not a bug 😅

[RoyTeddy]

LGTM and tested 👍
Conditional formatting works:

• in attachements ✅
• in body ✅

[liamlynch-data]

Found an issue but already there before: in the Body with custom value, if we specify a dataset name with numbers like {{ attachments.16rows.html_table }} -> beautiful error

...

it is a Python limitation. Attributes that start with a number, such as "123property" have to be expressed as {{customer['123property']}}

so works with {{ attachments['16rows'].html_table }} not a bug 😅

This one seems pre-existing, nothing to do with the change, and kind of as designed (or a limitation we can't do much about) - so we can totally ignore this can't we @RoyTeddy ?
(Thanks for the thorough testing though!)

[RoyTeddy]

Found an issue but already there before: in the Body with custom value, if we specify a dataset name with numbers like {{ attachments.16rows.html_table }} -> beautiful error

...

it is a Python limitation. Attributes that start with a number, such as "123property" have to be expressed as {{customer['123property']}}

so works with {{ attachments['16rows'].html_table }} not a bug 😅

This one seems pre-existing, nothing to do with the change, and kind of as designed (or a limitation we can't do much about) - so we can totally ignore this can't we @RoyTeddy ? (Thanks for the thorough testing though!)

@liamlynch-data yes we can ignore it 😉

[liamlynch-data]

Regression found: if Subject (column) or Body (column) is a datetime column we have this error: [18:05:25] [INFO] [dku.utils] - TypeError: unsupported operand type(s) for +: 'datetime.datetime' and 'str' Works with previous version 1.0.3 in release Version: 14.3.2 here https://released-design.qa.managedinstances.dkucloud-dev.com/projects/TEDDY_COPY/flow/

error caused by the input dataset which contains an empty row so no bug here 😅

Yes this is not a regression, v 1.0.3 of the plugin does this, e.g.:
https://staging-design.qa.managedinstances.dkucloud-dev.com/projects/SENDMAIL1/datasets/ergergerg/explore/

[MayeulRousselet]

Arriving late to the party, but I have a fundamental issue: If there are multiple recipient in the same row, and 1 CC/BCC, the CC/BCC will get spammed as he receive one email per recipient.

2 solutions:

• Cheapest: having a warning next to Recipient and CC/BCC
• Hardest, but better: havnig the possibility to send the same email to multiple recipient (need a DSS change)

[liamlynch-data]

Arriving late to the party, but I have a fundamental issue: If there are multiple recipient in the same row, and 1 CC/BCC, the CC/BCC will get spammed as he receive one email per recipient.

2 solutions:

• Cheapest: having a warning next to Recipient and CC/BCC
• Hardest, but better: havnig the possibility to send the same email to multiple recipient (need a DSS change)

I think we need to do the cheap option for now - @holstbone would you be able to add the warning? If you're not sure what needs doing I could send you a patch, but it should be simple

Re " the possibility to send the same email to multiple recipient" - we have a card already for this, but it's not scheduled:
https://app.shortcut.com/dataiku/story/233246/sendmail-add-an-option-so-if-multiple-senders-they-all-get-the-same-email

I'd also like to reiterate that we should really fix the following issue before we release with this (not necessarily in the same PR but before we release a version of the plugin):
https://app.shortcut.com/dataiku/story/298155/sendmail-plugin-wrong-column-type-for-email-fields-could-be-handled-better

[MayeulRousselet]

Arriving late to the party, but I have a fundamental issue: If there are multiple recipient in the same row, and 1 CC/BCC, the CC/BCC will get spammed as he receive one email per recipient.
2 solutions:

• Cheapest: having a warning next to Recipient and CC/BCC
• Hardest, but better: havnig the possibility to send the same email to multiple recipient (need a DSS change)

I think we need to do the cheap option for now - @holstbone would you be able to add the warning? If you're not sure what needs doing I could send you a patch, but it should be simple

Re " the possibility to send the same email to multiple recipient" - we have a card already for this, but it's not scheduled: https://app.shortcut.com/dataiku/story/233246/sendmail-add-an-option-so-if-multiple-senders-they-all-get-the-same-email

I'd also like to reiterate that we should really fix the following issue before we release with this (not necessarily in the same PR but before we release a version of the plugin): https://app.shortcut.com/dataiku/story/298155/sendmail-plugin-wrong-column-type-for-email-fields-could-be-handled-better

Logged as a blocker in SH @liamlynch-data

[holstbone]

Hi @MayeulRousselet and @liamlynch-data I have updated the descriptions for cc_column and bcc_column to include a note about duplicate data causing multiple emails.

Added text: "Note: Duplicate data will cause multiple emails."

Hopefully this addresses your concern!

It looks like this now:

[holstbone]

Hi @MayeulRousselet - I wanted to follow up on two items:

1.) Does my adjustment made on April 15th address your concern? See: #21 (comment)

2.) Regarding Shortcut #298155, did you actually test on this Pull Request (with the new code)? We did some better handling of types, but not certain if it specifically addresses the issue in the Shortcut. See: https://app.shortcut.com/dataiku/story/298155/sendmail-plugin-wrong-column-type-for-email-fields-could-be-handled-better

[liamlynch-data]

Hi @MayeulRousselet - I wanted to follow up on two items:

1.) Does my adjustment made on April 15th address your concern? See: #21 (comment)

2.) Regarding Shortcut #298155, did you actually test on this Pull Request (with the new code)? We did some better handling of types, but not certain if it specifically addresses the issue in the Shortcut. See: https://app.shortcut.com/dataiku/story/298155/sendmail-plugin-wrong-column-type-for-email-fields-could-be-handled-better

Re 2) - I raised this after testing this PR, so this is something we still need to fix. But I can do it after this one is merged, no worries.