Hi,
I want to automatically match data in an Excel file. The process is to merge 5 files : 1 Master file and 4 target files in one file.
The Master file has 2 columns : Reference and Label (designation).
Each target files has 2 columns : Label (designation) and CA TTC (turnover).
The 4 target files have to match the labels of the Master file. Their lines are displayed on the same line of the Master file.
It gives something like this.
I’ve been trying to do it through Levenshtein distance algorithm without success.
My files were created by hand and even though the familly order is the same (normally), labels are matching at 70% roughly, only.
Files have no primary key / reference, except for the Master file.
We have to match the lines with the labels only.
That is why I thought of using an AI Agent in n8n.
Here is the workflow. The first part before the Merge node is the whole process to format and agregate datas into one global files. It works fine. No need to check it.
I struggle having the AI Agent to work (after the node « Code - agrégation de tous les fichiers dans un seul »). Its tasks are very simple : insert lines in the target files columns to make the target lines match the Master file’s lines. That’s it. Nothing more.
I always get errors like « not a valid Json files » or the process gives me a file with correct labels but empty data, or other types of errors…
Could you help me fix this, please ?
Thanks in advance.