Download PALANTIR DATA ENGINEERING CERTIFICATION EXAM QUESTIONS AND VERIFIED ANSWERS and more Exams Engineering in PDF only on Docsity!
PALANTIR DATA ENGINEERING
CERTIFICATION EXAM QUESTIONS
AND VERIFIED ANSWERS ALREADY
GRADED A+ LATEST VERSION
"Which of the following practices are essential when implementing pipelines that back ontology objects and links in Foundry? Select two. Aligning pipeline logic with the ontology's entity and relationship definitions. Using only default transformation settings without customization. Avoiding documentation to keep the pipeline simple. Manually verifying each pipeline run for consistency. Ensuring that data transformations preserve the integrity of semantic relationships. Implementing error handling to manage discrepancies between data sources and ontology requirements. - CORRECT ANSWER=> Aligning pipeline logic with the ontology's entity and relationship definitions. Ensuring that data transformations preserve the integrity of semantic relationships." "You are assigned to maintain a critical data pipeline in Foundry that has been experiencing intermittent failures. To ensure timely resolution and support, which of the following support structures should you establish? Implement a ticketing system for tracking support requests and resolutions. Create detailed documentation outlining common issues and troubleshooting steps. Set up automated alerting for pipeline failures and performance issues. Restrict access to the pipeline only to senior data engineers. - CORRECT ANSWER=> Implement a ticketing system for tracking support requests and resolutions. Create detailed documentation outlining common issues and troubleshooting steps. Set up automated alerting for pipeline failures and performance issues."
"A data scientist wants to leverage their existing Jupyter notebooks within Palantir AIP for data analysis without switching to a different interface. Which feature of Palantir AIP should they utilize to achieve this? REST Interfaces Virtual Tables Palantir HyperAuto Pipelines Code Workspaces - CORRECT ANSWER=> Code Workspaces" "What are the kinetic elements in the Palantir Ontology? Objects, Properties, Links Actions, Functions Semantics, Interfaces Object Types, Link Types - CORRECT ANSWER=> Actions, Functions" "Which Linux operating system version is specifically recommended for hosting a Foundry agent? Ubuntu 18. Fedora 34 Debian 10 Red Hat Enterprise Linux 8 - CORRECT ANSWER=> Red Hat Enterprise Linux 8" "What actions are performed when the ModelOutput.publish() method is called in Foundry's Code Repositories? Select two: It serializes the model using the ModelAdapter.save() method. It initializes the model adapter with the fresh model. It runs the model inference. It creates a new model version. - CORRECT ANSWER=> It serializes the model using the ModelAdapter.save() method. It creates a new model version." "Which of the following statements correctly describes the behavior of the FileSystem.open() method in Foundry Transforms? it allows random access to any part of that file
add a dependency on 'transforms-media' in your code repository create media sets directly in the Python code use the @ media_set_input decorator to specify media sets - CORRECT ANSWER=> add a dependency on 'transforms-media' in your code repository" "Which of the following features can you utilize within Foundry's debugger panel while debugging a Python transform? Select three. Running PySpark commands in the console Previewing intermediate dataframes at breakpoints Automatically fixing variable values Changing the programming language of the transform on the fly Editing the source code directly from the debugger Navigating frames to examine variables - CORRECT ANSWER=> Running PySpark commands in the console Previewing intermediate dataframes at breakpoints Navigating frames to examine variables" "Which type of pipeline in Foundry typically has the lowest compute cost? all have similar compute costs streaming batch incremental - CORRECT ANSWER=> incremental" "In Foundry, which schema field type requires specifying both precision and scale parameters? ARRAY DECIMAL DATE STRING - CORRECT ANSWER=> DECIMAL" "Which of the following are stages included in the condaPackRun task used in CI checks for a Python repository in Foundry? Select three. Run unit tests using PyTest Upload artifacts to a remote server Download and extract all packages in the solved environment
Compile the Python source code Link packages into the environment Verify package contents - CORRECT ANSWER=> Download and extract all packages in the solved environment Link packages into the environment Verify package contents" "32. Which of the following Python libraries is NOT recommended for training models in Foundry's Code Repositories? scikit-learn SparkML PyTorch TensorFlow - CORRECT ANSWER=> SparkML" "33. Which of the following are recommended practices for refactoring complex logical operations in PySpark transformations? Chain multiple 'filter()' and 'withColumn()' calls in a single line Extract complex logic into separate functions. Use deeply nested parentheses to encapsulate logical operations. Group logic into named variables. Keep logic expressions inside the same code block to 3 expressions at most. Duplicate code for better readability. - CORRECT ANSWER=> Extract complex logic into separate functions. Group logic into named variables. Keep logic expressions inside the same code block to 3 expressions at most." "36. You are developing a Transform in Foundry that processes input dataframes using PySpark and needs to output multiple datasets based on different filters. Which decorator should you use to define this Transform? @ transform_df @ transform_pandas @ transform_file @ transform - CORRECT ANSWER=> @ transform"
When you want to maintain a detailed commit history with merge commits. When there are no additional changes on the target branch and you want a linear commit history. - CORRECT ANSWER=> When there are no additional changes on the target branch and you want a linear commit history." "47. You want to leverage distributed processing in Foundry Transforms to handle files of varying sizes efficiently. Which Spark configuration properties should you adjust to control the partitioning of the FileStatus DataFrame? Select two. spark.executor.cores spark.executor.memory spark.sql.files.openCostInBytes spark.driver.memory spark.sql.files.maxPartitionBytes - CORRECT ANSWER=> spark.sql.files.openCostInBytes spark.sql.files.maxPartitionBytes" "48. In a Foundry Pipeline, you need to generate multiple output datasets from a single input dataset by filtering based on different criteria. Which feature of the Transforms API allows you to accomplish this efficiently? Multiple-output Transforms Transform logic level versioning (TLLV) Transform generation using for-loops TransformContext injection - CORRECT ANSWER=> Multiple-output Transforms" "50. Which of the following are recommended practices when performing join operations in PySpark according to the style guide? Select two. Use dataframe aliases to manage column references Specify the join type explicitly, even if it's the default Prefer left joins and avoid right joins Always use right joins instead of left joins Use .dropDuplicates() to handle join explosions Duplicate column names to avoid ambiguity - CORRECT ANSWER=> Specify the join type explicitly, even if it's the default Prefer left joins and avoid right joins"
"54. Which of the following health checks are recommended to install on input datasets of a Foundry data pipeline? Select three. Build Status Check Schema Check Data Freshness Build Duration Check Sync Status Check Time Since Last Updated (TSLU) - CORRECT ANSWER=> Build Status Check Schema Check Time Since Last Updated (TSLU)" "57. Which of the following are recommended practices for handling shared datasets used across multiple pipelines in Foundry? Select two. Increase the complexity of your pipeline setup by integrating shared datasets directly into multiple pipelines Allow each pipeline to build the shared dataset independently based on their own schedules Treat the shared dataset as an input in only one pipeline and ignore it in others Enable multiple schedules to trigger builds for the shared dataset simultaneously Create a new pipeline dedicated to building the shared dataset and have other pipelines treat it as an input Use Data Connection syncs to manage shared datasets - CORRECT ANSWER=> Treat the shared dataset as an input in only one pipeline and ignore it in others Create a new pipeline dedicated to building the shared dataset and have other pipelines treat it as an input" "58. You need to completely replace the existing data in a dataset with a new batch of data. Which type of transaction should you perform in Foundry? APPEND DELETE UPDATE SNAPSHOT - CORRECT ANSWER=> SNAPSHOT" "59. You have created a new repository named 'Data_Processor' for your shared Python library in Foundry. According to Conda's naming conventions, how will this repository name be published as a Conda package?
Assign one active developer per individual branch. Have multiple users share the same branch. Restrict branch creation to administrators only. Merge changes frequently to avoid conflicts. - CORRECT ANSWER=> Assign one active developer per individual branch." "8. To set up test coverage reporting in your Python repository using PyTest, which of the following steps should you perform? Select two. Create a pytest.ini file with coverage options. Install the coverage package using pip separately. Configure the build.gradle file to include coverage tasks. Add 'pytest-cov' to the test requirements in meta.yml. - CORRECT ANSWER=> Add 'pytest-cov' to the test requirements in meta.yml." "10. What are the necessary steps to configure an incremental batch sync for a JDBC connection in Foundry? Select three. Disable the preview functionality before configuring incremental sync. Set the transaction type to APPEND. Modify the SQL query to include a WHERE clause with the incremental column using the wildcard '?' Set the initial value of the incremental column to zero regardless of previous syncs. Use the Overwrite transaction type to ensure data consistency. Enable the Incremental option and configure the incremental state. - CORRECT ANSWER=> Set the transaction type to APPEND. Modify the SQL query to include a WHERE clause with the incremental column using the wildcard '?' Enable the Incremental option and configure the incremental state." "12. Which of the following practices help in minimizing breaking changes when modifying dataset schemas? Select two. Modifying existing columns to repurpose them for new data types. Avoiding any changes to the schema to prevent breaking changes. Creating new columns instead of modifying or deleting existing ones. Deleting old columns immediately after adding new ones.
Announcing deprecation of old columns and providing instructions to data consumers. - CORRECT ANSWER=> Creating new columns instead of modifying or deleting existing ones. Announcing deprecation of old columns and providing instructions to data consumers." "16. You have set up a post-condition Data Expectation on the output of a transform in Foundry. If this expectation fails during a build, what will occur? The build is automatically aborted, and the output is not written. The input dataset is aborted to prevent issues. The build is resumed with a warning. The failed expectation is ignored, and the build continues. - CORRECT ANSWER=> The build is automatically aborted, and the output is not written." "17. In a Foundry Transform, you need to perform random access to a file to read specific lines multiple times. Given that FileSystem.open() does not support random access, what is a recommended workaround? Split the file into smaller chunks and process them separately. Buffer the entire file into memory using io.StringIO or a temporary file. Enable random access by configuring FileSystem to support seek and tell. Use multiple FileSystem.open() calls to access different parts of the file. - CORRECT ANSWER=> Buffer the entire file into memory using io.StringIO or a temporary file." "24. When defining a Transform with multiple outlets, how should you write the compute function to utilize the filtered DataFrame only once for generating all outputs? Filter the DataFrame separately for each output within the compute function. Leverage the TransformContext to manage DataFrame filtering. Filter the DataFrame once and assign it to a variable, then use that variable to generate each output. Use multiple compute functions, each handling a different output. - CORRECT ANSWER=> Filter the DataFrame once and assign it to a variable, then use that variable to generate each output." "26. You are debugging a Python transform in Foundry and find that a breakpoint within an internal library is colored grey. What should you do to effectively debug this location? Enable internal libraries debugging in the debugger settings. Ignore the breakpoint and proceed with execution.
@file_transform @transform_files @transform - CORRECT ANSWER=> @transform" "37. Which of the following actions will trigger re-resolution of Conda lock files in Foundry's Code Repositories? Select three. Running Task Runner. Upgrading to a newer template version. Changing the list of packages in the meta.yaml file. Adding a new branch to the repository. Modifying the build.gradle file. Deleting the hidden Conda lock files. - CORRECT ANSWER=> Upgrading to a newer template version. Changing the list of packages in the meta.yaml file. Deleting the hidden Conda lock files." "38. You are developing a Transform in Foundry that needs to read only JSON files from an input dataset for further processing. Which method and parameter should you use to list these files efficiently? filesystem.list('.json') filesystem.open('.json') filesystem.read_files('.json') filesystem.ls(glob='.json') - CORRECT ANSWER=> filesystem.ls(glob='*.json')" "43. Which of the following are considered product types as defined in the release process? Use-case product Transform product Workflow product Ontology product Feature product Schema product - CORRECT ANSWER=> Use-case product Ontology product" "45. What determines the starting point for calculating a dataset view in Foundry?
The latest SNAPSHOT transaction before that point in time. The earliest transaction in the dataset. The first DELETE transaction. The latest APPEND transaction. - CORRECT ANSWER=> The latest SNAPSHOT transaction before that point in time." "46. Which decorator must be used when defining a Python transform that utilizes media sets in Foundry? @transform_media @transform @media_transform @media_set - CORRECT ANSWER=> @transform" "56. Which of the following steps are necessary for publishing a trained model in Foundry's Code Repositories? Select two. Use SparkML for training. Call ModelOutput.publish() to save the model. Author a model adapter. Write a Python transform to train the model. - CORRECT ANSWER=> Call ModelOutput.publish() to save the model. Author a model adapter." "58. You are setting up a PySpark DataFrame transformation in Foundry and want to ensure that the output DataFrame adheres to a specific schema. What method should you primarily use at the beginning of your transformation to define the schema contract? collect() show() select() withColumn() - CORRECT ANSWER=> select()" "60. You have completed developing a new feature on a feature branch derived from 'dev'. After merging into 'dev', what is the next step to integrate the feature into the production 'master' branch? Create a new feature branch from 'master' and merge it into 'dev'.
Select the dataset and filter the list of related artifacts to include object types. - CORRECT ANSWER=> Select the object type from the search results to add it to your data lineage graph. Use the Search Foundry tool in the right sidebar to find the desired object type." "18. You initiated a build on a feature branch with a fallback chain of feature → master, where dataset A is on master. During the build, two jobs are executed serially: the first job writes to dataset B on the feature branch, and the second job writes to dataset C on the feature branch. What will be the state of dataset A after the build? Dataset A on the master branch is updated with new data. Dataset A on the feature branch is updated with new data. Both feature and master branches of dataset A are updated. Dataset A remains unchanged. - CORRECT ANSWER=> Dataset A remains unchanged." "20. Which of the following file formats is recommended to store unstructured data within a Foundry dataset? Parquet Text JSON Avro - CORRECT ANSWER=> Text" "24. How can you disable specific PyLint messages, such as 'missing-module-docstring', in your Python project within Foundry? Use command-line arguments when running PyTest to disable the messages. Remove the associated code that triggers the messages. Edit the build.gradle file to exclude these messages. Modify the src/.pylintrc file to disable the specific messages. - CORRECT ANSWER=> Modify the src/.pylintrc file to disable the specific messages." "33. Which of the following features are available under the Details view in Foundry's Dataset Preview? Select three. Adding custom metadata fields Viewing and downloading dataset files Comparing datasets Monitoring real-time data streams
Editing the dataset schema Scheduling data syncs - CORRECT ANSWER=> Adding custom metadata fields Viewing and downloading dataset files Editing the dataset schema" "34. You have transitioned a data pipeline to maintenance mode and need to ensure it continues to meet user requirements. What should you define first before starting the maintenance process? The user access permissions for the pipeline The cost of maintaining the pipeline The pipeline's data scope and delivery expectations The technical architecture of upstream systems - CORRECT ANSWER=> The pipeline's data scope and delivery expectations" "52. You are developing a Transform within the 'Data Cleaning Project' in Foundry. Your Transform requires access to a dataset owned by the 'Customer Data Project.' According to Project references guidelines, what action should you take to include the 'Customer Data Project' dataset as an input for your Transform? Export the 'Customer Data Project' dataset to the 'Data Cleaning Project'. Add a Project reference to the 'Customer Data Project' dataset. Update the code repository's language packages to include the 'Customer Data Project'. Directly reference the dataset without any additional configuration. - CORRECT ANSWER=> Add a Project reference to the 'Customer Data Project' dataset." "55. You are tasked with writing a model to the output dataset in Foundry using the pickle module. Which mode should you use when opening the file with FileSystem.open()? 'r' 'wb' 'rb' 'w' - CORRECT ANSWER=> 'wb'" "57. You are tasked with setting up a new transform in Palantir Foundry that utilizes Palantir's OpenAI GPT-4 language model to analyze customer feedback. Which of the following steps should you perform first?
"11. You are managing a data pipeline in Foundry that processes incoming sales data, transforms it, and outputs a final dataset for reporting. To effectively monitor this pipeline, where should you install Schema Checks to detect any unexpected changes in the data structure? Only on intermediate datasets Only on the input datasets On both input and output datasets Only on the output datasets - CORRECT ANSWER=> On both input and output datasets" "19. To add support for Palantir-provided language models in your Foundry transforms, which library should you install from the Code Repository's library search panel? pandas numpy palantir_models scipy - CORRECT ANSWER=> palantir_models" "25. Which Gradle plugin should be applied to enable the Spark anti-pattern linter in your Python project within Foundry? 'com.palantir.conda.pylint' 'com.palantir.conda.pep8' 'com.palantir.transforms.lang.pytest-defaults' 'com.palantir.transforms.lang.antipattern-linter' - CORRECT ANSWER=> 'com.palantir.transforms.lang.antipattern-linter'" "26. Which of the following functions are performed by Foundry builds concerning branches? Select two. Resolving job inputs and outputs with respect to the build branch and fallback branches. Creating new branches for each build. Automatically deleting unused branches after a build. Compiling the build graph by collecting JobSpecs from branches. Merging changes from multiple branches into the build branch. Ensuring that builds modify all dataset branches. - CORRECT ANSWER=> Resolving job inputs and outputs with respect to the build branch and fallback branches.
Compiling the build graph by collecting JobSpecs from branches." "35. Which section within the Information panel of Foundry's Dataset Preview provides details such as the dataset's creation time, last update, and the users responsible for these actions? About Columns Schedules Data Preview - CORRECT ANSWER=> About" "37. Which of the following methods adhere to the recommended PySpark style when adding new columns to a DataFrame? Using withColumnRenamed to add new columns. Adding new columns directly without specifying the method. Using select with multiple expressions to add new columns. Using withColumn to add each new column individually. - CORRECT ANSWER=> Using withColumn to add each new column individually." "40. In the recommended branching strategy, what is the primary role of the 'master' branch? It is used to create short-lived feature branches. It integrates schema changes at specific cadences. It is the production branch and is sourced with production data. It serves as the staging branch for testing new features. - CORRECT ANSWER=> It is the production branch and is sourced with production data." "42. When defining Transform logic level versioning (TLLV), which of the following factors are included in the default version string? Select three. The module where the Transform is defined The runtime environment configuration The names of all input datasets All functions within the Transform Any project dependencies All modules the Transform depends on - CORRECT ANSWER=> The module where the Transform is defined Any project dependencies
