HTML Image size issue in email

[dana]

Hi,

I have created different reports, each has the same table, the same width but different height.

In office outlook 365, the images have the same width.

But in  Outlook classic, the images have different width.

Can it be resolved?

Thanks in advance!

[Lech_Miszkiewicz]

Hi Dana,

I am sure you are already aware of this but this is not really NPrinting issue but Microsoft issue and the way their email clients render html.

The answer on microsoft community about image rendering should give you hint how to tackle this:

https://learn.microsoft.com/en-us/answers/questions/4679132/how-to-fix-size-of-email-signature-displ...

Bottom line is - there is no NPrinting property we can use to fix this immidiately and fixing such solution requires proper testign on proper example. 

Lastly - I can see that you are using just straight tables to produce this. Using images for me is totally last resort so I would suggest to use HTML and data table to build native HTML table embeded into your email. You can easily achieve nice formatting, better scalability and 100% better look with native HTML tables. I would have never gone with images as that always look odd (blurry, out of shape, different font size/style and out of proportions). Maybe you can reconsider.

cheers

[dana]

Hi @Lech_Miszkiewicz,

Thanks for the reference and the tips..

I have defined the same width and height for all of the tables, but since there is a significant height difference, it results in a long "tail" of empty space in the shorter tables.

Re HTML table:

I'm not familiar with HTML writing, and not sure how can this be achieved, since the straight table has calculated dimension, and use of Pick for the expressions. Can you please elaborate?

Thank you!

[Lech_Miszkiewicz · Accepted solution]

Hi @dana 

building html table will require you to learn basic html tags and learn how styling can be applied as the same principles apply to images as to tables regarding supported styling of objects in different emails clients. 

Regarding actual process there is not much:

https://help.qlik.com/en-US/nprinting/February2025/Content/NPrinting/HTMLReports/HTML-Reports-Contai... - this is general overview

and this is general formatting concept:

https://help.qlik.com/en-US/nprinting/February2025/Content/NPrinting/HTMLReports/HTML-formatting.htm

You just need to spend time learning it and practicing - there is no copy/paste solution for this. It requires prep on Qlik Sense/QlikVIew app side as well as when building template but the final result in my opinion is always worth it!

cheers

[Lech_Miszkiewicz]

Ohh - just following up on your concern - you should be just fine with pick() expression and calculated dimension in this case but you may have to apply text() on measures in order to ensure those are the formats applied (in case they are not coming through by themselves). Note that keep source format should be disabled and html encode should be used to apply formatting based on records coming from your Qlik app as eplained here:https://community.qlik.com/t5/Qlik-NPrinting/Nprinting-HTML-How-change-background-color-based-on-the...

cheers

[dana]

Hi @Lech_Miszkiewicz ,

I checked the HTML table option you offered.

I didn't see an option to expand my table.

According to ChatGPT, it is due to calculated dimension or conditional show.

• NPrinting metadata relies on “static” field definitions it can read from the Sense app.

• A calculated dimension or conditionally hidden column doesn’t look like a stable field to NPrinting, so it won’t appear in the template editor tree.

I have both:
Hidden future months,  4 total lines on top.

Indeed, I could expand a simple straight table.
So, looks like I'll have to find another solution..

Just thought I should mention this..

[Lech_Miszkiewicz]

Hi @dana 

we need to strongly say one thing:
Chat gpt IS USELESS when it comes to Nprinting! It makes things up, hallucinates more than you can imagine and creates more confusion than good. I am helping on this community regularly and amount of nonsense questions asked by people mislead by chat gpt is so big that it worries me as i can see people are actually listening to it instead of checking facts. My suggestion is check existing community posts and dont rely on chat gpt because it does not say full truth but only cherry picks contextual words without understanding how Nprinting works. 

Now, there are always ways of doing things. Hiding and showing columns can also be achieved as i described in this blog post: https://nprintingadventures.com/2020/04/24/how-to-build-html-pivot-table-in-nprinting/

To have individual columns available you need to use straight table and dont use conditional hide/show column.

yes it may sound complicated but it can be achieved with workarounds like formatting using css and html and conditional expressions inside of always visible columns.

The process of designing such report starts with:

• defining required output
• defining things which are dynamic/moving/conditional in structure of your report and focus on how you can achieve them
• looking for equivalent features in css/html which support all required mail clients (note that <script> tags for example will never work in email clients
• If none of the options will work on regular straight table grom Qlik considering workarounds you can always produce whole html script as shown in my pivot table blog post which can be very much extended and improved

cheers

[dana]

Hi @Lech_Miszkiewicz ,

I would divide it to 2 parts: diagnose and solution of the issue.
I think that in my case, the diagnose of chat GPT was correct. Do you agree?
Chat GPT offered a few workarounds as well, but didn't publish them because I didn't check them yet.

Thanks for proposing useful workarounds and a link to your post 🙂

[Lech_Miszkiewicz]

Hi @dana 

I agree to respectively disagree 😜 or should i say: lets not dwell on chat gpt too much. 

bottom line for me is that I would not trust it 100%  and if I can’t trust it I don't waste my time on it. 

re Nprinting my approach always is to make sure i use supported approach and i try to find supported workaround to get final result as good as possible.