Parsing website search history JSON data in PowerBI - json

Question

I am doing some analytics on a certain website, and I'm trying to get people's search terms to an easier format. They're in JSON format, and I've already parsed the column several times to get other relevant data. The problem is that when trying to parse the keywords/search data, each search word goes into its own column, and with tens of thousands of searches it becomes a problem. Hard to explain, so I will add a picture that illustrates the problem.
The "Ai O" and "JAHAHASHAS" are searches that I did on the website, rest are hidden, but the list goes on forever
https://i.imgur.com/5IXTXn1.png
Here is an example of a person searching "business model" in the JSON
"{""sort"": ""lastUpdated"", ""limit"": 5, ""fields"": {""keywords"": {""business model"": true}}, ""offset"": 0}"
Tried to parse the keywords column anyways, but creating thousands of new columns doesn't really work

Answer 1

I added some extra records to your json file, I hope that it matches what you have to work with in reality. In the future, you'll get faster and better results if you're willing to share a bit more concrete information about your project. In most cases, it's very difficult to provide meaningful advice if we can't reasonably recreate your situation.
Anyway, I wanted the solution to work against multiple search terms, so I tweaked the JSON you provided to look like this:
{ "Searches": [
{"sort": "lastUpdated", "limit": 5, "fields": {"keywords": {"business model": true}}, "offset": 0},
{"sort": "lastUpdated", "limit": 5, "fields": {"keywords": {"Alpha Beta": true}}, "offset": 0},
{"sort": "lastUpdated", "limit": 5, "fields": {"keywords": {"Gamma Delta Epsilon": true}}, "offset": 0}
]}
and then provided the following Power Query (taken from the advanced editor)
let
Source = Json.Document(File.Contents("C:\Users\XXXXXX\Desktop\foo.json")),
#"Converted to Table" = Table.FromList(Source[Searches], Splitter.SplitByNothing(), null, null, ExtraValues.Error),
#"Expanded Column1" = Table.ExpandRecordColumn(#"Converted to Table", "Column1", {"fields"}, {"Fields"}),
#"Expanded Fields" = Table.ExpandRecordColumn(#"Expanded Column1", "Fields", {"keywords"}, {"keywords"}),
#"Added Custom" = Table.AddColumn(#"Expanded Fields", "keywords_text", each Record.FieldNames([keywords])),
#"Extracted Values" = Table.TransformColumns(#"Added Custom", {"keywords_text", each Text.Combine(List.Transform(_, Text.From), " "), type text}),
#"Added Custom1" = Table.AddColumn(#"Extracted Values", "keywords_text_list", each Text.Split([keywords_text], " ")),
Custom1 = List.Combine( #"Added Custom1"[keywords_text_list])
in
Custom1
After loading the data from file, I extracted the records (searches, that is) from the Json.
Next, having converted the data to a table, I expanded twice to dive down from Records to Fields, and Fields to keywords. Then I added a custom column and used the M Query 'Record.Fieldnames' function to generate a list on every fieldnames in our keyword collections. Keep in mind, each list here only has one element, even though there might be spaces in that text.
Then I extracted values from those lists
and followed that up by splitting those strings back into lists -- this time using space as a delimiter.
As a final step, I combined the multiple lists into a single list of all our keywords as single elements.
So, that's my approach. Hope it helps.