fix typo toml->yaml and add caution

docs/website/docs/walkthroughs/adjust-a-schema.md

@@ -28,7 +28,7 @@ dlt.pipeline(
 )
 ```
 
-Following folder structure in project root folder will be created:
+The following folder structure in the project root folder will be created:
 
 ```
 schemas
@@ -46,11 +46,10 @@ import_schema_path="schemas/import"
 ## 2. Run the pipeline to see the schemas
 
 To see the schemas, you must run your pipeline again. The `schemas` and `import`/`export`
-directories will be created. In each directory, you'll see a `yaml` file with a file
-`chess.schema.toml`.
+directories will be created. In each directory, you'll see a `yaml` file (e.g. `chess.schema.yaml`).
 
 Look at the export schema (in the export folder): this is the schema that got inferred from the data
-and was used to load it into the destination (i.e `duckdb`).
+and was used to load it into the destination (e.g. `duckdb`).
 
 ## 3. Make changes in import schema
 
@@ -59,20 +58,36 @@ hints that were explicitly declared in the `chess` source. You'll use this schem
 modifications, typically by pasting relevant snippets from your export schema and modifying them.
 You should keep the import schema as simple as possible and let `dlt` do the rest.
 
-> 💡 How importing a schema works:
->
-> 1. When a new pipeline is created and the source function is extracted for the first time, a new
->    schema is added to pipeline. This schema is created out of global hints and resource hints
->    present in the source extractor function.
-> 1. Every such new schema will be saved to the `import` folder (if it does not exist there already)
->    and used as the initial version for all future pipeline runs.
-> 1. Once a schema is present in `import` folder, **it is writable by the user only**.
-> 1. Any change to the schemas in that folder are detected and propagated to the pipeline
->    automatically on the next run. It means that after an user update, the schema in `import`
->    folder reverts all the automatic updates from the data.
-
-In next steps we'll experiment a lot, you will be warned to set `full_refresh=True` in the
-`dlt.pipeline` until we are done experimenting.
+💡 How importing a schema works:
+
+1. When a new pipeline is created and the source function is extracted for the first time, a new
+   schema is added to the pipeline. This schema is created out of global hints and resource hints
+   present in the source extractor function.
+1. Every such new schema will be saved to the `import` folder (if it does not exist there already)
+   and used as the initial version for all future pipeline runs.
+1. Once a schema is present in `import` folder, **it is writable by the user only**.
+1. Any changes to the schemas in that folder are detected and propagated to the pipeline
+   automatically on the next run. It means that after a user update, the schema in `import`
+   folder reverts all the automatic updates from the data.
+
+In next steps we'll experiment a lot, you will be warned to set `full_refresh=True` until we are done experimenting.
+
+:::caution
+`dlt` will **not modify** tables after they are created.
+So if you have a `yaml` file, and you change it (e.g. change a data type or add a hint),
+then you need to **delete the dataset**
+or set `full_refresh=True`:
+```python
+dlt.pipeline(
+    import_schema_path="schemas/import",
+    export_schema_path="schemas/export",
+    pipeline_name="chess_pipeline",
+    destination='duckdb',
+    dataset_name="games_data",
+    full_refresh=True,
+)
+```
+:::
 
 ### Change the data type
 
@@ -97,7 +112,7 @@ players_games:
       data_type: timestamp
 ```
 
-Run the pipeline script again and make sure that the change is visible in export schema. Then,
+Run the pipeline script again and make sure that the change is visible in the export schema. Then,
 [launch the Streamlit app](../dlt-ecosystem/visualizations/exploring-the-data.md) to see the changed data.
 
 :::note
@@ -122,7 +137,7 @@ white__aid:
   data_type: text
 ```
 
-For some reason you'd rather deal with a single JSON (or struct) column. Just declare the `white`
+For some reason, you'd rather deal with a single JSON (or struct) column. Just declare the `white`
 column as `complex`, which will instruct `dlt` not to flatten it (or not convert into child table in
 case of a list). Do the same with `black` column:
 
@@ -166,5 +181,5 @@ players_games:
 
 ## 4. Keep your import schema
 
-Just add and push the import folder to git. It will be used automatically when cloned. Alternatively
+Just add and push the import folder to git. It will be used automatically when cloned. Alternatively,
 [bundle such schema with your source](../general-usage/schema.md#attaching-schemas-to-sources).