I feel this took me too much time to figure out, so I am posting a quick showcase to maybe save future makers some time.
Background: Our company gets email notifications from our bank whenever there are collections/payments received. The email always includes an HTML table, but it can be in a different order or contain extra fields.
Solution: Using a mailhook, we can setup a make scenario that instant triggers every time an email is sent to our inbox we can ingest it in make and send the relevant data to where it needs to get to.
Showcase: See the screenshots below, but what I wanted to share with the community is how I used the âText parserâ to âget content from an HTML tableâ and then iterate and aggregate that data to make it usable in our scenario.
The interesting finding was the Get Content from HTML Table worked really well, but the data was Array to Collection to Array. So I was struggling to get the Key/Value pairings with the get and map functions (maybe someone can share on this thread how to do that more easily).
So the solution was using the âtext aggregatorâ module, and specifically the group by with the text to aggregate, I could effectively create a Key:Value pairing that I was used to having.
Ultimately I end up with very easy to work with bundles, where each bundle has a key and a value. From here I ended up putting in a data store, then I can use âget a recordâ calls to the data store to pull out the exact data I want to pass to a Google Sheet (and Slack).
I am always eager to learn and continue to grow my make skills, so if anyone else in the community has a more efficient or nice way to solve for this, I would be happy to hear about it. In the meantime, I hope someone else finds this helpful, as I couldnât find much on the community regarding the âGet content from HTML tableâ and how to actually transform that to a usable array or bundle of data.
Thanks, thatâs good to know! I faced a similar situation, I got turned off by the complex ouput, gave up and turned to Zapierâs text parser. The tool is quite powerful and easy to use with plain text templates, which can be refined via iterations.
I will try to give your approach a try on the next use case.
This is really helpful and. In my case, I had to use Array Aggregator to create a single bundle since Iâm getting multiple bundles at once. Hope this will helpful to someone
A quick update, as my company is Enterprise user of Make.com, we do have access to create custom functions.
So the actual way I handle this type of situation now is with a couple of custom functions (written in Javascript in Make). My favorite is this âfindBetweenWordsâ function, here is the code below. I pass it the email text, write the start word and end word, and the function returns everything between. I understand not everyone has access to enterprise version of Make, but the custom functions can be extremely powerful. (Also note I am a real amateur at javascript so if you see any errors in this code feel free to let me know, but so far it has worked as intended).
function findBetweenWords(emailText, startWords, endWords) {
// Find the start and end positions
const startPos = emailText.indexOf(startWords);
const endPos = emailText.indexOf(endWords, startPos + startWords.length);
// Check if start and end positions are valid
if (startPos === -1 || endPos === -1 || endPos <= startPos) {
return null;
}
// Extract the text between start and end positions
const textBetween = emailText.slice(startPos + startWords.length, endPos).trim();
// Check if the extracted text contains only blank spaces
if (!textBetween.replace(/\s/g, '').length) {
return null;
}
return textBetween;
}
Thank you @Joseph_Accountant your original post was extremely helpful. However, as it only works for a single key/value pair, Iâm wondering if anyone has a solution for a table with three or more columns? Unless Iâm missing something, there is no way to get at objects in a second-level nested array that are in positions other than 1 and 2.
Hi @Barnaby_Marshall - thank you for the reply, and you are correct that my original solution is not helpful when dealing with three or more columns.
Maybe you can look at using the âMatch Pattern (Advanced)â within the Text Parser module? Some of the free AI models are pretty good at writing the REGEX code to extract data, but be warned that REGEX has itâs shortcomings and will require a lot of unit testing if your relying on it to process a lot of data.
Iâve been trying for hours, but I couldnât figure out how to pull the âtest-dataâ data in the screenshot, and the Make team didnât respond to my ticket.
Happy to point you in the right direction. In general I think you need to learn a bit more about how arrays work and specifically how arrays in Make work.
A few options come to mind.
Get & Map Array functions. See this great post linked below for some more detail.
If that value will always be in the same place you can also hard code the array level in Make. See screenshot below as a small sample, here I am hardcoding the 3rd item in the array at the first level, then the 2nd item in the array on the 2nd level, and ultimately getting back the âdescriptorâ field.
Use the Iterator module to iterate the various levels of the array, this is probably the most user friendly but the least efficient, but if you havenât tried it before it may help conceptualize how the data is actually being stored (another way to do this is âdownload the output bundleâ and paste it in a code editor that can format JSON, you will understand the structure better and it will make working with the data in Make less frustrating).
I hope this helps get you on the right path, and it may feel discouraging to spend hours trying to get the answer, but you will learn a lot along the way and the next time you have a similar problem you will breeze right through it.
