How can I preserve and merge binary PDF data for renaming in n8n?

Below is some guidance on why your binary data may be “lost” when using the Merge node and how to work around this.

---

What’s Happening

When you merge two inputs, one branch (your HTTP Request node) contains the binary data (your PDF file) while the other branch (e.g. a node that adds a new filename) likely contains only JSON data. By default the Merge node will combine the JSON properties from both inputs but won’t automatically merge binary properties if only one branch provides them. In your case, the binary file from the HTTP Request is being dropped (or not passed along) during the merge step.

---

Recommended Solutions

There are a couple of ways to address this:

1. Rename Without Merging
   Instead of merging two branches, try updating (renaming) the binary property directly. In many cases you can use a Set node (or even a Function/Code node) immediately after the HTTP Request node to modify its binary property. For example, if your HTTP Request returns a binary property under “data,” you can set a new filename by doing something like this in a Set node:

   JSONCopied
   {
     "binary": {
       "data": {
         "fileName": "my-renamed-file.pdf"
       }
     }
   }

   In n8n you’d add a field (e.g. “data.fileName”) as a string with your desired name—using an expression if dynamic naming is needed—so that when you later use the Google Drive node, the file is uploaded with the new name.

2. Merge by Position (or “Keep All”)
   If you really need to pull in additional JSON data from another branch, then make sure you choose a merge mode that preserves the binary data. For example, when using the Merge node you can choose “Merge By Position” so that the first item (which carries the binary property) is paired with your additional JSON (such as the new filename). If you use “Merge by Fields” make sure that both branches carry the field you are matching on, otherwise the merge may result in your binary data not being attached.

   Note that if one branch doesn’t contain any binary properties (it only adds a filename), the merge may simply not combine with the binary carried on the other branch. In these cases, it is often simpler to modify the binary property from the HTTP Request node itself using a Set or Function node—as in solution 1.

---

In Summary

• Why the Error?
  The Merge node is dropping the binary file because only one branch (your HTTP Request) has binary data; the branch that “renames” the file only provides JSON.
• How to Fix?
  Either rename directly on the HTTP Request output using a Set node (thus preserving the binary properties) or use a merge option that ensures the binary data is retained (which requires that both branches provide the expected fields).

For most users, the simplest solution is to avoid the merge and add a node right after the HTTP Request that updates the binary metadata (e.g. the file’s name) before passing it to your Google Drive node.

---

Give one of these methods a try and adjust your workflow accordingly. If you need further clarification on setting or merging binary data, feel free to ask!