# How to Search For Specific Words In All Fabric Notebooks In A Workspace

My personal Fabric trial tenant is a mess. I create items, test them and don’t name them properly. For example, I have a workspace with 226 notebooks named `Notebook 1` , `Notebook 2` etc 🙄 I am sure I am not alone of this crime. I needed to find a notebook I created a few months ago that contained a specific code logic. Fabric global search only searches for item names and not item content. So here is what I came up with:

```python
import sempy.fabric as fabric 
import json

def extract_notebook_cells(notebook_name, workspace):

    return [{'cell_type': cell.get('cell_type'), 'source': ''.join(cell.get('source', '')).replace('\n', ' ')} 
            for cell in json.loads(notebookutils.notebook.getDefinition(notebook_name, workspaceId=workspace)).get('cells', [])]

def search_notebooks(word, workspace=None):
    if workspace is None: #workspace id of the notebook if not given
        workspace = fabric.get_notebook_workspace_id() 
    else:
        workspace = fabric.resolve_workspace_id(workspace) #get workspace id
    
    notebooks = [nb['displayName'] for nb in notebookutils.notebook.list(workspace)]
    matching_notebooks = []
    print(f"Workspace: {fabric.resolve_workspace_name(workspace)}")
    for notebook_name in notebooks:
        try:
            cells = extract_notebook_cells(notebook_name=notebook_name, workspace=workspace)
            if any(word.lower() in str(cell['source']).lower() for cell in cells):
                matching_notebooks.append(notebook_name)
        except Exception as e:
            print(f"Error {notebook_name}: {str(e)}")
    
    return matching_notebooks

search_notebooks("Cache")
```

**Logic:**

* Get a list of notebooks in a workspace
    
* For each notebook, get the definition, strip everything and get only cell/markdown blocks
    
* Search for the keyword
    
* Depending on the number of notebooks and notebook content, this can take a while. You can certainly speed it up with some optimization but this is sufficient for me. You can also use your own regex pattern for more advanced search.
    

## UPDATE : Aug 26, 2025

Semantic Link Labs v0.11.3 now has a [`search_notebooks`](https://github.com/microsoft/semantic-link-labs/wiki/Code-Examples#search-notebooks) function to search a keyword in a single notebook or all notebooks in a list of workspaces.

![](https://cdn.hashnode.com/res/hashnode/image/upload/v1756219937457/6348689c-57ff-4d4f-8444-ab1ceb775800.png align="center")