Merge pull request #242960 from santiagxf/santiagxf/mlflow-tracking

Update how-to-track-experiments-mlflow.md

Author: prmerger-automator[bot]

articles/machine-learning/how-to-track-experiments-mlflow.md

@@ -135,28 +135,38 @@ All metrics and parameters are also returned when querying runs. However, for me
 
 By default, experiments are ordered descending by `start_time`, which is the time the experiment was queue in Azure Machine Learning. However, you can change this default by using the parameter `order_by`.
 
-* Order runs by the attribute `start_time`:
+* Order runs by attributes, like `start_time`:
 
     ```python
     mlflow.search_runs(experiment_ids=[ "1234-5678-90AB-CDEFG" ],
                        order_by=["attributes.start_time DESC"])
     ```
 
-* Order and retrieve the last run:
+* Order runs and limit results. The following example returns the last single run in the experiment:
 
     ```python
     mlflow.search_runs(experiment_ids=[ "1234-5678-90AB-CDEFG" ], 
                        max_results=1, order_by=["attributes.start_time DESC"])
     ```
 
+* Order runs by the attribute `duration`:
+
+    ```python
+    mlflow.search_runs(experiment_ids=[ "1234-5678-90AB-CDEFG" ], 
+                       order_by=["attributes.duration DESC"])
+    ```
+
+    > [!TIP]
+    > `attributes.duration` is not present in MLflow OSS, but provided in Azure Machine Learning for convenience.
+
 * Order runs by metric's values:
 
     ```python
     mlflow.search_runs(experiment_ids=[ "1234-5678-90AB-CDEFG" ]).sort_values("metrics.accuracy", ascending=False)
     ```
 
     > [!WARNING]
-    > Using `order_by` with expressions containing `metrics.*` in the parameter `order_by` is not supported by the moment. Please use `order_values` method from Pandas as shown in the next example.
+    > Using `order_by` with expressions containing `metrics.*`, `params.*`, or `tags.*` in the parameter `order_by` is not supported by the moment. Please use `order_values` method from Pandas as shown in the example.
 
 
 ### Filtering runs
@@ -213,7 +223,18 @@ You can also look for a run with a specific combination in the hyperparameters u
 
     > [!TIP]
     > Notice that for the key `attributes`, values should always be strings and hence encoded between quotes.
-  
+
+* Search runs taking longer than one hour:
+
+    ```python
+    duration = 360 * 1000 # duration is in milliseconds
+    mlflow.search_runs(experiment_ids=[ "1234-5678-90AB-CDEFG" ], 
+                       filter_string=f"attributes.duration > '{duration}'")
+    ```
+
+    > [!TIP]
+    > `attributes.duration` is not present in MLflow OSS, but provided in Azure Machine Learning for convenience.
+    
 * Search runs having the ID in a given set:
 
     ```python