# Learning Palantir Foundry 🚀 "Sales reps see only the rows for their assigned customers" — achieved on a single dataset, without spawning a copy per department. That is what Restricted Views (row-level security) deliver. 📌 Title and Feature URL Title: 制限付きビュー(行レベルセキュリティ) URL: 📝 Overview Restricted Views implement granular row-level access control. A restricted view is built on top of a backing dataset and lets different users see different subsets of the same underlying data based on defined permissions. This removes the need to copy datasets per team: you keep one dataset as the source of truth shared across the company while splitting visibility at the row level. 🔧 How It Works Restricted views operate through policies containing rules that determine row visibility. - Policies evaluate the viewing user's attributes, column names from the backing dataset, and specific values (strings, Booleans, numbers, arrays) to decide which rows are shown. - When referencing users, groups, or organizations, you must use the unique identifier (UUID) in both the policy column and the policy definition — names alone will not work. - In marking-backed views, the upstream dataset holds a STRING ARRAY column of Marking IDs, and each row is visible only to users with the required markings. - A restricted view is built on top of a backing dataset and cannot be used as an input for transforms. - Experimental branching support allows adding and merging restricted view policy changes. 🛠 Practical Usage - Add a column to the backing dataset that drives row access (for example, assigned branch or organization ID). - Define a policy that matches that column against user attributes to build the row-level filter. - Save restricted views in a separate Project from the source datasets to keep access management clean. - For markings, attach an array of required Marking IDs per row to control visibility. 🎯 Use Cases - Limit sales representatives to viewing customers at their assigned branch. - Separate records on a shared table by department or organization. - Disclose differently classified records only to users holding the required markings. ⚠️ Caveats - Restricted views cannot serve as transform inputs, so they cannot be plugged directly into downstream pipeline processing. - Users, groups, and organizations must be referenced by UUID; name-based references do not work. - Merging policy changes via branching is experimental and may not be universally available. #PalantirFoundry# #DataGovernance#

Run inference over millions of records — free of SQL, and without your data ever leaving Snowflake. Here's distributed batch inference at scale ⚙️ Title: Batch Inference at Scale URL: ⚙️ Overview A capability that runs distributed inference workloads on Snowpark Container Services (SPCS) with Ray as the execution framework. Inference runs as a dedicated distributed workload, supporting both traditional models and LLMs, consolidating complex operations into a single API call. ❓ Challenges Solved Many customers, especially those migrating from non-SQL systems, need batch inference decoupled from SQL. ・This is especially true for files and unstructured data at large scale ・Rearchitecting workflows around SQL-first patterns is a heavy burden 💡 Methodology & How It Works ・The input DataFrame is materialized and written to a stage as Parquet files ・A job is provisioned on SPCS; the primary node initializes as the Ray head and replicas join as workers ・Each worker reads staged data, performs inference independently, and writes results to an output stage ・Unified API: a single run_batch() call handles both structured and unstructured data ・Multimodal support (images, audio, video); workers load weights once and reuse across batches; JobSpec controls workers and GPU allocation 🌍 Use Cases ・Nightly summarization of millions of support tickets ・Product catalog enrichment via image-to-text generation ・Information extraction from scanned PDFs, audio transcription and labeling, video classification and description BatchInferenceTask integrates with Snowflake Tasks for DAG automation, and all processing stays inside Snowflake — running large-scale inference while preserving data governance. #Snowflake# #BatchInference#