Databricks-DA-Assoc Analyzing Queries — All Questions With Answers

Question 1

An analyst notices that a specific query is slow because it performs a full table scan instead of using the intended index. Which command should the analyst check to ensure the optimizer has the necessary information to choose the correct plan?

Accepted Answer

ANALYZE TABLE.. In Databricks, the command to collect statistics for the cost-based optimizer is 'ANALYZE TABLE table_name COMPUTE STATISTICS'. However, the stem's premise regarding 'indexes' is misleading for Databricks, as it relies on Z-Ordering or Data Skipping rather than traditional indexes. Furthermore, the command 'ANALYZE TABLE' is often deprecated in favor of 'ANALYZE TABLE table_name COMPUTE STATISTICS FOR COLUMNS'.

Answer

OPTIMIZE TABLE.

Answer

VACUUM TABLE.

Answer

REFRESH TABLE.

Question 2

A data analyst is investigating a slow Databricks SQL query that joins two large Delta tables. The query profile shows a SortMergeJoin with two Exchange nodes, and each Exchange shuffles over 1 TB of data. The analyst wants to reduce shuffle. Which approach is most likely to improve performance?

Accepted Answer

Use bucketing on the join key for both tables and ensure the number of buckets matches.. Bucketing both tables on the join key with the same number of buckets allows Spark to perform a sort-merge join without exchanging data across the network. This eliminates the two Exchange nodes and the associated 1 TB shuffles, significantly improving performance. Other options either do not eliminate shuffle or are not applicable when both tables are large.

Answer

Pre-partition both tables on the join key using Delta Lake partitioning.

Answer

Rewrite the query to use a broadcast join hint for the smaller table.

Answer

Enable adaptive query execution and set spark.sql.adaptive.coalescePartitions.enabled to true.

Question 3

A data analyst uses a Databricks SQL dashboard that refreshes every 15 minutes. One query reads from a large Delta table, and the analyst notices in the Query Profile that the scan operator reads far more bytes than the rows the final result returns, even though the WHERE clause filters on a low-cardinality status column. The analyst wants to reduce bytes scanned without changing the result. Which action is most appropriate?

Accepted Answer

Use a materialized view that pre-aggregates or pre-filters the data the dashboard needs. Because the filter targets a low-cardinality column, file-level skipping and partitioning both leave large amounts of data to scan. A materialized view precomputes the filtered or aggregated result so the recurring dashboard query reads a compact table, preserving the result while sharply reducing bytes scanned on each refresh.

Answer

Partition the Delta table by the status column and rely on partition pruning

Answer

Add a ZORDER BY clause on the status column when optimizing the Delta table

Answer

Enable Delta Lake data skipping statistics and filter on a high-selectivity column instead

Question 4

A data analyst runs a query that includes a subquery in the WHERE clause: SELECT * FROM orders WHERE customer_id IN (SELECT customer_id FROM customers WHERE region = 'North'). The query is slow. The analyst wants to rewrite the query to improve performance. Which approach is most likely to be more efficient?

Accepted Answer

Rewrite the query as an INNER JOIN between orders and customers on customer_id, with a filter on region.. Rewriting the IN subquery as an INNER JOIN with a filter on region makes the join explicit, enabling Spark to apply optimizations like broadcast join if the customers table is small. This often results in a more efficient execution plan than a subquery, which may be evaluated separately.

Answer

Use a LEFT SEMI JOIN instead of the IN subquery.

Answer

Add a broadcast hint to the orders table to force a broadcast join.

Answer

Replace the IN subquery with an EXISTS correlated subquery.

Question 5

A data analyst is examining the Databricks SQL Query Profile for a query that is slow due to a join. The analyst wants to identify whether the join is causing a shuffle and whether data skew is present. Which two metrics should the analyst examine? (Choose two.)

Accepted Answer

The number of rows and bytes shuffled per stage.. Shuffle metrics (rows and bytes) directly show whether data is being exchanged and how much, confirming a shuffle-based join. Peak memory usage per task helps identify skew: if some tasks use much more memory, they are likely processing more data due to skewed keys. Together, these metrics diagnose the two issues.

Answer

The duration of each task and the distribution of task durations.

Answer

The query plan's node types and their order.

Answer

The number of partitions in the table.

Question 6

A data analyst runs a SELECT * query on a Delta table with 10,000 partitions, but only 5 partitions match the WHERE clause on a partition column. The analyst observes that the query still scans all partitions. Which action should the analyst take to ensure partition pruning occurs?

Accepted Answer

Use a filter on the partition column in the WHERE clause.. Partition pruning is a critical optimization in Databricks SQL that relies on the query filter referencing the partition column. Without such a filter, Spark must scan all partitions, leading to unnecessary I/O. The correct action is to ensure the WHERE clause includes a condition on the partition column, allowing the optimizer to skip irrelevant partitions.

Answer

Increase the number of shuffle partitions.

Answer

Add a Z-ORDER BY clause on the partition column.

Answer

Run OPTIMIZE on the table to compact small files.

Question 7

A data analyst runs a Databricks SQL query against a Delta table and wants to inspect the detailed logical and physical execution plan, including statistics used by the cost-based optimizer, to understand why a filter was not pushed down. The analyst has access to the query's statement ID from the Query History. Which approach provides the most detailed execution plan information?

Accepted Answer

Prefix the query with EXPLAIN EXTENDED and review the output.. To understand why a filter was not pushed down, the analyst needs the optimized logical and physical plans along with statistics. EXPLAIN EXTENDED outputs these plans and exposes optimizer decisions such as predicate pushdown, making it the right tool. Runtime-only views like the Query Profile or Spark UI show execution behavior but not the plan transformations that determine pushdown.

Answer

Enable the Spark UI and inspect the SQL tab's DAG visualization for the query.

Answer

Open the Query History, select the query, and view the Query Profile's operator-level metrics.

Answer

Run DESCRIBE DETAIL on the Delta table referenced in the query.

Question 8

A data analyst is troubleshooting a slow-running SQL query against a massive Delta table in Databricks. The query frequently scans the entire table despite filtering on a high-cardinality timestamp column. Which approach will most effectively reduce the data scanned by eliminating full-table reads?

Accepted Answer

Run an OPTIMIZE command with a ZORDER BY clause on the timestamp column to co-locate related data and improve data skipping.. Z-Ordering co-locates related data based on specified columns, significantly improving data skipping for queries with equality or range filters. When combined with correct partitioning, it minimizes the amount of data scanned from cloud storage, drastically reducing query latency and execution costs in Databricks environments.

Answer

Increase the cluster size to a driver instance with more memory to hold the entire uncompressed Delta table in cache.

Answer

Execute a VACUUM command with a retention threshold of zero hours to purge old data files immediately.

Answer

Convert the Delta table format to standard Parquet files to take advantage of native Apache Spark partitioning.

Question 9

A data analyst needs to inspect the logical and physical execution plans of a slow Spark SQL query to understand how filters and joins are being evaluated. Which command should the analyst execute in a Databricks notebook cell?

Accepted Answer

EXPLAIN EXTENDED SELECT * FROM table_name WHERE condition;. The EXPLAIN command in Spark SQL outputs the execution plans, including parsed logical plan, analyzed logical plan, optimized logical plan, and physical plan. Analyzing these plans helps data analysts identify inefficient join strategies, missing predicate pushdown, and unnecessary full table scans.

Answer

DESCRIBE EXTENDED table_name;

Answer

SHOW QUERY PLAN FOR SELECT * FROM table_name WHERE condition;

Answer

ANALYZE TABLE table_name COMPUTE STATISTICS FOR COLUMNS;

Question 10

A data analyst notices that a query involving a large join between two tables is consistently slow. The analyst suspects that one of the tables is significantly skewed. Which tool in the Databricks SQL query profile is most effective for confirming this skew?

Accepted Answer

The SQL Query Profile 'Metrics' tab.. The Query Profile provides visual metrics to identify bottlenecks. By examining the 'Max vs Median' row count metrics per task, an analyst can see if specific executors are processing significantly more data than others, indicating skew. Understanding data distribution is critical in Databricks, as skew often leads to 'straggler' tasks that delay entire query execution, regardless of overall cluster size or compute power.

Answer

The Query History list view.

Answer

The Cluster Usage dashboard.

Answer

The Table details metadata panel.

Question 11

An analyst is optimizing a query that performs multiple aggregations on a Delta table. Which TWO actions can the analyst take to improve query performance via the SQL editor?

Accepted Answer

Use Z-Ordering on columns frequently used in WHERE filters.. Improving performance requires reducing the amount of data read. Using partition pruning allows the engine to skip irrelevant files, while Z-Ordering improves data skipping by clustering related information together. These techniques are fundamental for Databricks SQL performance, as they minimize I/O overhead. Mastering these adjustments allows analysts to write highly efficient code that scales effectively with growing datasets, reducing both latency and costs for the organization.

Answer

Increase the cluster size to 'Large' for every query execution.

Answer

Force a Broadcast Join for all small tables.

Answer

Convert all tables to Parquet format.

Question 12

Which of the following is the primary purpose of examining the 'Query Profile' in Databricks SQL?

Accepted Answer

To visualize the execution steps and identify bottlenecks.. The Query Profile is the central diagnostic tool for Databricks SQL. It provides a visual representation of how a query was executed, including the time spent in each stage and where bottlenecks occur. Mastering this tool is vital for data analysts because it transforms abstract performance issues into actionable insights, enabling them to reduce query latency, optimize costs, and effectively communicate performance bottlenecks to data engineers for backend architecture improvements.

Answer

To change the permissions of the underlying table.

Answer

To automatically rewrite SQL for better performance.

Answer

To view the raw logs of the cluster driver.

Question 13

When analyzing query performance, what does a high 'Spill to Disk' metric indicate?

Accepted Answer

The cluster has insufficient memory for the operation.. Spilling to disk occurs when a query's operations, such as sorts or joins, exceed the available memory in the cluster's executors. This is a major performance bottleneck because disk I/O is exponentially slower than memory. Recognizing this indicator is crucial for analysts because it signals that the current cluster size or query logic needs adjustment to fit the workload comfortably within RAM, significantly improving query speed and stability.

Answer

The query is successfully using the cache.

Answer

The data is highly compressed.

Answer

The query is finished and writing results to the table.

Question 14

An analyst is reviewing a slow query and identifies that the 'FileScan' stage is taking most of the time. Which THREE factors could be causing this inefficiency?

Accepted Answer

The table contains a high number of small files.. Slow FileScans usually point to I/O-related issues. If too many small files are present, the metadata overhead becomes significant. Without proper pruning, the system scans unnecessary data. Z-Ordering or partitioning issues cause the engine to read more data than required. Addressing these factors is vital for analysts, as they directly impact the 'Data Skipping' efficiency of the Delta Lake engine, which is the cornerstone of high-performance analytics in Databricks.

Answer

The query uses a cross join.

Answer

The query is using a User-Defined Function (UDF).

Question 15

Why should an analyst use the 'EXPLAIN' command before running a complex SQL query on a very large dataset?

Accepted Answer

To verify the query plan and identify inefficiencies.. The EXPLAIN command provides the logical and physical execution plan without actually running the query. For analysts, this is a proactive safety and performance measure. It allows them to spot inefficient patterns, such as unexpected cross joins or massive data shuffles, before consuming expensive compute resources. This practice is essential for cost control and ensuring that queries are optimized for the scale of the underlying production data.

Answer

To bypass the authentication requirement for queries.

Answer

To view the projected cost in US Dollars.

Answer

To automatically apply Z-Ordering to the table.

Question 16

When analyzing query execution metrics in Databricks, what does 'Task Duration' represent?

Accepted Answer

The time elapsed for a single core to process a task.. Task Duration measures the time taken by individual executor nodes to perform their assigned portion of the work. It is distinct from total query time because it highlights parallel execution efficiency. Understanding this is vital because if one task is much longer than others, it indicates data skew or uneven distribution, helping analysts identify the root cause of performance bottlenecks that are hidden within the overall query execution time.

Answer

The time taken to compile the SQL query.

Answer

The total time the cluster was running.

Answer

The latency added by network overhead only.

Question 17

Which feature in Databricks SQL allows an analyst to view the history and performance metrics of previously executed queries?

Accepted Answer

The Query History.. The Query History interface is the primary tool for reviewing past executions. It allows analysts to search, filter, and inspect the performance of queries run across the workspace. This is important for identifying long-running queries, diagnosing failures, and comparing performance over time, which supports the iterative process of optimizing data models and SQL code to ensure consistent, reliable, and performant data delivery for the organization.

Answer

The Data Explorer.

Answer

The Cluster Metrics tab.

Answer

The Workspace Audit Logs.

Question 18

What is the primary benefit of using a 'Materialized View' in Databricks SQL for frequent, expensive queries?

Accepted Answer

It stores precomputed results to improve read performance.. Materialized views precompute query results, allowing for faster read performance by storing the output rather than recalculating it on every request. This is critical for data analysts working with large, complex datasets, as it significantly reduces latency and compute costs for repetitive dashboards or reports. By shifting the workload from query time to refresh time, analysts can provide consistent and fast performance for their business stakeholders.

Answer

It bypasses the need for Unity Catalog permissions.

Answer

It automatically scales the cluster to infinite size.

Answer

It eliminates the need for any data partitioning.

Question 19

An analyst is troubleshooting a query that hangs indefinitely during a join. Which TWO metrics in the Query Profile should the analyst examine to diagnose the issue?

Accepted Answer

Shuffle Read/Write bytes.. When a join hangs, it is often due to massive data movement (shuffle) or extreme data skew. Monitoring the shuffle size helps determine if the network or I/O is saturated. Checking the 'Max vs Median' row count helps identify skew. These metrics are the most reliable indicators of join failure, enabling analysts to decide whether to adjust join keys, use broadcast hints, or address underlying data distribution issues.

Answer

The number of users logged into the workspace.

Answer

The total number of tables in the schema.

Answer

The total size of the database logs.

Question 20

When reviewing a query profile, an analyst sees a 'Broadcast Hash Join'. What does this tell the analyst about the data being joined?

Accepted Answer

One of the tables is small enough to fit in memory.. A Broadcast Hash Join indicates that one side of the join is small enough to be sent to all executor nodes. This is an extremely efficient join type as it eliminates shuffling. For analysts, recognizing this confirms that the optimizer has successfully identified a small table, which is a positive performance indicator. If a join is expected to be small but isn't broadcast, the analyst may need to check table statistics.

Answer

The join is extremely memory-intensive.

Answer

The query is performing a cross join.

Answer

The data is skewed and needs re-partitioning.

Databricks-DA-Assoc Analyzing Queries — All Questions With Answers

An analyst notices that a specific query is slow because it performs a full table scan instead of using the intended index. Which command should the analyst check to ensure the optimizer has the necessary information to choose the correct plan?

A data analyst is investigating a slow Databricks SQL query that joins two large Delta tables. The query profile shows a SortMergeJoin with two Exchange nodes, and each Exchange shuffles over 1 TB of data. The analyst wants to reduce shuffle. Which approach is most likely to improve performance?

A data analyst is examining the Databricks SQL Query Profile for a query that is slow due to a join. The analyst wants to identify whether the join is causing a shuffle and whether data skew is present. Which two metrics should the analyst examine? (Choose two.)

A data analyst runs a SELECT * query on a Delta table with 10,000 partitions, but only 5 partitions match the WHERE clause on a partition column. The analyst observes that the query still scans all partitions. Which action should the analyst take to ensure partition pruning occurs?

A data analyst is troubleshooting a slow-running SQL query against a massive Delta table in Databricks. The query frequently scans the entire table despite filtering on a high-cardinality timestamp column. Which approach will most effectively reduce the data scanned by eliminating full-table reads?

A data analyst needs to inspect the logical and physical execution plans of a slow Spark SQL query to understand how filters and joins are being evaluated. Which command should the analyst execute in a Databricks notebook cell?

A data analyst notices that a query involving a large join between two tables is consistently slow. The analyst suspects that one of the tables is significantly skewed. Which tool in the Databricks SQL query profile is most effective for confirming this skew?

An analyst is optimizing a query that performs multiple aggregations on a Delta table. Which TWO actions can the analyst take to improve query performance via the SQL editor?

Which of the following is the primary purpose of examining the 'Query Profile' in Databricks SQL?

When analyzing query performance, what does a high 'Spill to Disk' metric indicate?

An analyst is reviewing a slow query and identifies that the 'FileScan' stage is taking most of the time. Which THREE factors could be causing this inefficiency?

Why should an analyst use the 'EXPLAIN' command before running a complex SQL query on a very large dataset?

When analyzing query execution metrics in Databricks, what does 'Task Duration' represent?

Which feature in Databricks SQL allows an analyst to view the history and performance metrics of previously executed queries?

What is the primary benefit of using a 'Materialized View' in Databricks SQL for frequent, expensive queries?

An analyst is troubleshooting a query that hangs indefinitely during a join. Which TWO metrics in the Query Profile should the analyst examine to diagnose the issue?

When reviewing a query profile, an analyst sees a 'Broadcast Hash Join'. What does this tell the analyst about the data being joined?

A data analyst runs a query against a large Delta table partitioned by date. The query filters on a timestamp column rather than the date partition column. How does Databricks handle this query execution?

An analyst executes a query that joins a large fact table with a small dimension table in Databricks. The analyst wants to force the Catalyst optimizer to broadcast the small dimension table to avoid an expensive shuffle join. Which standard Spark SQL hint should be included in the query text?

A data analyst is investigating why a specific Spark SQL query takes an exceptionally long time to complete and exhibits signs of severe data skew. Which TWO metrics or behaviors in the Databricks Spark UI typically indicate that data skew is impacting the query? (Choose two)

An analyst is writing a query in Databricks and needs to reference a temporary view that is scoped only to their current interactive notebook session. Which command or syntax ensures the view is automatically dropped when the session ends?

A data analyst is querying a large Delta table containing billions of rows of clickstream data. The table is frequently queried using a timestamp column. To optimize range queries on this timestamp, which Delta Lake feature should the table builder implement?

Refer to the exhibit. An analyst examines the physical execution plan of a Spark SQL join query. Based on the provided plan output, which optimization strategy was automatically applied by Adaptive Query Execution (AQE)?

Exhibit

Refer to the exhibit. A data analyst reviews a physical execution plan where a table is scanned and shuffled multiple times in succession during a self-join operation. What underlying anti-pattern in the query construction likely caused this redundant exchange?

Exhibit

A data analyst notices that a query involving a join between a large table and a small lookup table is performing poorly. The analyst wants to optimize the join performance without changing the underlying data. Which technique should they apply?

When analyzing the execution plan of a query in Databricks SQL, what does a 'BroadcastHashJoin' node typically indicate?

An analyst is reviewing a slow-running query and wants to identify potential bottlenecks. Which TWO of the following metrics in the Databricks SQL query profile are most indicative of inefficient data distribution?

Refer to the exhibit. The execution plan shows a SortMergeJoin with two Exchange nodes. What is the primary cause of the performance impact here?

Exhibit

An analyst is optimizing a dashboard that queries a Delta table partitioned by 'event_date'. The query filters on 'event_date' and 'user_id', but is still slow. What is the most effective next step to improve query speed?

A data analyst is analyzing a slow-running Databricks SQL query. The analyst suspects that the query is suffering from excessive shuffling. Which two actions should the analyst take to reduce shuffle and improve performance? (Choose two.)

A data analyst is reviewing the Query Profile in Databricks SQL for a query that spills data to disk during a sort operation. Which two metrics should the analyst examine to confirm and diagnose the spill? (Choose two.)

A data analyst is investigating a query that uses a window function with PARTITION BY and ORDER BY, and the query is slow. The analyst suspects that data skew is causing some partitions to be much larger than others. Which approach should the analyst take to diagnose the skew in the Query Profile?

Databricks-DA-Assoc Analyzing Queries — All Questions With Answers

An analyst notices that a specific query is slow because it performs a full table scan instead of using the intended index. Which command should the analyst check to ensure the optimizer has the necessary information to choose the correct plan?

A data analyst is investigating a slow Databricks SQL query that joins two large Delta tables. The query profile shows a SortMergeJoin with two Exchange nodes, and each Exchange shuffles over 1 TB of data. The analyst wants to reduce shuffle. Which approach is most likely to improve performance?

A data analyst is examining the Databricks SQL Query Profile for a query that is slow due to a join. The analyst wants to identify whether the join is causing a shuffle and whether data skew is present. Which two metrics should the analyst examine? (Choose two.)

A data analyst runs a SELECT * query on a Delta table with 10,000 partitions, but only 5 partitions match the WHERE clause on a partition column. The analyst observes that the query still scans all partitions. Which action should the analyst take to ensure partition pruning occurs?

A data analyst is troubleshooting a slow-running SQL query against a massive Delta table in Databricks. The query frequently scans the entire table despite filtering on a high-cardinality timestamp column. Which approach will most effectively reduce the data scanned by eliminating full-table reads?

A data analyst needs to inspect the logical and physical execution plans of a slow Spark SQL query to understand how filters and joins are being evaluated. Which command should the analyst execute in a Databricks notebook cell?

A data analyst notices that a query involving a large join between two tables is consistently slow. The analyst suspects that one of the tables is significantly skewed. Which tool in the Databricks SQL query profile is most effective for confirming this skew?

An analyst is optimizing a query that performs multiple aggregations on a Delta table. Which TWO actions can the analyst take to improve query performance via the SQL editor?

Which of the following is the primary purpose of examining the 'Query Profile' in Databricks SQL?

When analyzing query performance, what does a high 'Spill to Disk' metric indicate?

An analyst is reviewing a slow query and identifies that the 'FileScan' stage is taking most of the time. Which THREE factors could be causing this inefficiency?

Why should an analyst use the 'EXPLAIN' command before running a complex SQL query on a very large dataset?

When analyzing query execution metrics in Databricks, what does 'Task Duration' represent?

Which feature in Databricks SQL allows an analyst to view the history and performance metrics of previously executed queries?

What is the primary benefit of using a 'Materialized View' in Databricks SQL for frequent, expensive queries?

An analyst is troubleshooting a query that hangs indefinitely during a join. Which TWO metrics in the Query Profile should the analyst examine to diagnose the issue?

When reviewing a query profile, an analyst sees a 'Broadcast Hash Join'. What does this tell the analyst about the data being joined?

A data analyst runs a query against a large Delta table partitioned by date. The query filters on a timestamp column rather than the date partition column. How does Databricks handle this query execution?

An analyst executes a query that joins a large fact table with a small dimension table in Databricks. The analyst wants to force the Catalyst optimizer to broadcast the small dimension table to avoid an expensive shuffle join. Which standard Spark SQL hint should be included in the query text?

A data analyst is investigating why a specific Spark SQL query takes an exceptionally long time to complete and exhibits signs of severe data skew. Which TWO metrics or behaviors in the Databricks Spark UI typically indicate that data skew is impacting the query? (Choose two)

An analyst is writing a query in Databricks and needs to reference a temporary view that is scoped only to their current interactive notebook session. Which command or syntax ensures the view is automatically dropped when the session ends?

A data analyst is querying a large Delta table containing billions of rows of clickstream data. The table is frequently queried using a timestamp column. To optimize range queries on this timestamp, which Delta Lake feature should the table builder implement?

Refer to the exhibit. An analyst examines the physical execution plan of a Spark SQL join query. Based on the provided plan output, which optimization strategy was automatically applied by Adaptive Query Execution (AQE)?

Exhibit

Refer to the exhibit. A data analyst reviews a physical execution plan where a table is scanned and shuffled multiple times in succession during a self-join operation. What underlying anti-pattern in the query construction likely caused this redundant exchange?

Exhibit

A data analyst notices that a query involving a join between a large table and a small lookup table is performing poorly. The analyst wants to optimize the join performance without changing the underlying data. Which technique should they apply?

When analyzing the execution plan of a query in Databricks SQL, what does a 'BroadcastHashJoin' node typically indicate?

An analyst is reviewing a slow-running query and wants to identify potential bottlenecks. Which TWO of the following metrics in the Databricks SQL query profile are most indicative of inefficient data distribution?

Refer to the exhibit. The execution plan shows a SortMergeJoin with two Exchange nodes. What is the primary cause of the performance impact here?

Exhibit

An analyst is optimizing a dashboard that queries a Delta table partitioned by 'event_date'. The query filters on 'event_date' and 'user_id', but is still slow. What is the most effective next step to improve query speed?

A data analyst is analyzing a slow-running Databricks SQL query. The analyst suspects that the query is suffering from excessive shuffling. Which two actions should the analyst take to reduce shuffle and improve performance? (Choose two.)

A data analyst is reviewing the Query Profile in Databricks SQL for a query that spills data to disk during a sort operation. Which two metrics should the analyst examine to confirm and diagnose the spill? (Choose two.)

A data analyst is investigating a query that uses a window function with PARTITION BY and ORDER BY, and the query is slow. The analyst suspects that data skew is causing some partitions to be much larger than others. Which approach should the analyst take to diagnose the skew in the Query Profile?