SQL/PySpark interview roadmap
A roadmap post lays out a structured set of interview practice problems—dozens of SQL exercises plus Python and PySpark challenges—focused on window functions, joins and performance tuning common in analytics and quant interviews. The collection is presented as a progressive path for data‑engineer and analytics preparation. (x.com)
SQL and PySpark interview prep is increasingly being packaged as a step-by-step drill set: start with basic queries, then move to joins, window functions, Python coding, and Spark tuning. (x.com) The roadmap in this post is organized as a progression rather than a random question bank, with SQL exercises first and Python and PySpark challenges layered on top. The topics highlighted in the post — joins, window functions, and performance work — match the core areas covered in common Spark and data-engineering practice materials. (x.com) (github.com) A join is the database operation that combines rows from two tables using a shared key, and it is one of the first filters interviewers use to test whether a candidate can work across related datasets. Window functions keep each row visible while calculating ranks, running totals, or previous-row values, which is why Apache Spark documents them separately from ordinary group-by aggregates. (spark.apache.org 1) (spark.apache.org 2) PySpark is the Python interface to Apache Spark, which runs data processing across many machines instead of one laptop. Spark’s own Python documentation defines `pyspark.sql.Window` as the utility for building those per-row window calculations inside DataFrames. (spark.apache.org 1) (spark.apache.org 2) Performance tuning is the part of the interview where syntax stops being enough and candidates have to explain why a query runs slowly. Apache Spark’s current documentation lists caching, partition changes, join strategy selection, statistics, and Adaptive Query Execution among the main levers for speeding up DataFrame and Spark SQL workloads. (spark.apache.org) That focus reflects how analytics and data-engineering interviews have shifted from “write a SELECT statement” to “write it and explain the execution plan.” Amazon Web Services’ Spark tuning guidance says window functions, group-by operations, and broadcast joins can have different runtime tradeoffs depending on data size and column cardinality. (docs.aws.amazon.com) The same pattern shows up in broader interview repositories. One GitHub handbook for data engineers lists 100-plus PySpark coding patterns, more than 20 SQL pattern collections, and 15-plus Spark performance topics alongside system design and cloud tools. (github.com) For candidates, the practical takeaway is that interview prep is being framed less as memorizing isolated answers and more as building fluency across three layers: SQL logic, Python problem-solving, and distributed-computing judgment. That is the path this roadmap is selling — and it lines up with the way Spark’s own documentation separates syntax, APIs, and optimization. (x.com) (spark.apache.org 1) (spark.apache.org 2)
