Finally, it seems I found a way, maybe not the best but I did it as follows.

I first inspected the webpage and noted down the following table details div.horizontal-scroll-wrapper :

Which I reported in the HTML Extractor:

Then it gave me the table extraction as follows in an attribute:

<div class="horizontal-scroll-wrapper">

<table class="table table-bordered table-narrow">

<thead>

<tr>

<th scope="col" width="30%">Mesures</th>

<th class="text-center" scope="col">Débit<br/><small class="text-muted">09.07.2021 21:40</small><br/><small class="text-muted">m³/s</small></th>

<th class="text-center" scope="col">Niveau d'eau<br/><small class="text-muted">09.07.2021 21:40</small><br/><small class="text-muted">m</small></th>

<th class="text-center" scope="col">Température<br/><small class="text-muted">09.07.2021 21:40</small><br/><small class="text-muted">°C</small></th>

</tr>

</thead>

<tbody>

<tr>

<th scope="row">Dernière mesure</th>

<td class="text-center">739</td>

<td class="text-center">336.05</td>

<td class="text-center">17.5</td>

</tr>

<tr>

<th scope="row">Moyenne 24h</th>

<td class="text-center">828</td>

<td class="text-center">336.30</td>

<td class="text-center">16.5</td>

</tr>

<tr>

<th scope="row">Maximum 24h</th>

<td class="text-center">897</td>

<td class="text-center">336.50</td>

<td class="text-center">17.5</td>

</tr>

</tbody>

</table>

</div>

Then I added a second HTML Extractor to parse this new attribute, and used a list :

Then I had all the records and renamed the list with an attribute manager:

I add this in case it might help someone else. It might not be the best method though.

Hi @topotoma​ ,

How about this, use a FeatureReader to read in the HTML table in the Single Output Port mode, then use a BulkAttributeRenamer to remove the datetime/temperature portion of the attribute name using Regex and a date stamp with a wildcard. 

@DateTimeFormat(@DateTimeNow(),%d\.%m\.%Y).*

From there you can attach an AttributeManager to expose the attributes and continue processing or just attach it to your writer :) 

@DateTimeFormat(@DateTimeNow(),%d\.%m\.%Y).*

Hi Crisatsafe,

That's pretty neat, I tried to do that but i didn't know the syntax for renaming the attributes nor did I had a clue how to get it right.

However if I use your fmw the table has missing values at the end:

Maybe that's because the time stamp has changed but I couldnt see why as your parameters don't interfer with it.

Nevertheless may thanks for your quick reply and guidance to an elegant solution.

I also need to keep the original timestamp from the webpage, it is an useful information to know the exact time of the river level measurements, so my version with the HTMLExtractor does that at the moment.

Thanks again and best regards.

Thomas

Ah perfect! Didn't see your update before posting that - glad you were able to get it all working  

Just re-ran it and also noticed the missing values and that appears to be because of the time difference between my machine (PST) and the website time (CEST) so I had to update it to the following to make up the time difference (9 hours in my case) as it had already rolled into July 10 on that site:

@DateTimeFormat(@DateTimeAdd(@DateTimeNow(),PT9H),%d\.%m\.%Y).*

BUT not important since you have it all working now ;) great solution!

Thank you Chrisatsafe, that explains it all.

It would have been great to have the possibility to copy the value of the portion of the attribute name containing the date/time and reuse it later as a parameter.

Maybe with a Schema Reader, then with another transformer to extract a certain number of characters from the text. (a bit like 'right' formula in excel).

I didn't figure out how to do it yet.

Anyway, many thanks for your help and have a good day.

Best regards,

Thomas

That is something that is possible using the SubstringExtractor or using a String Function like GetWord in the AttributeManager (or even a StringExtractor since we know the regex for it). The string functions will behave similar to all of the string functions in excel so you could certainly extract the info from the header and include it in an attribute if needed.