SQL GROUP BY: Avoiding Not a GROUP BY Expression Error

SQL is the language of data, but are you using its "hidden" logic? 🔍 Writing queries is one thing; understanding the engine is another. Here are 4 things about SQL that changed how I think about data: - The Execution Order Lie: We write SELECT first, but SQL executes it almost last. It starts with FROM and WHERE. This is why you can’t use a column alias in your filter—the engine hasn't "seen" the alias yet! - The NULL Trap: In SQL, NULL = NULL is False (technically Unknown). NULL is a state, not a value. If you use NOT IN on a list containing a NULL, your whole query might return zero results. - SARGable Queries: If you use a function on a column in your WHERE clause (like WHERE YEAR(date) = 2025), you might be killing your performance. It prevents the database from using indexes. Use a date range instead. - Window Functions > Group By: SUM() OVER() is often more powerful than a standard GROUP BY. It allows you to keep your row-level detail while adding aggregate context in the same view. SQL isn't just about getting the data; it’s about getting it efficiently. 🚀 What’s one SQL "gotcha" that caught you off guard when you first started? ⬇️ #SQL #DataAnalytics #DataEngineering #CodingTips #Database #PowerBI

📊 Day 6/100: THE "SELECT" COMMAND Yesterday, we talked about queries, the questions we ask our data Mastering how and when to use the SELECT command is crucial, non-negotiable and a game changer in your journey as an analyst. with that being Today, let’s dive into the most important SQL command which is the SELECT command If SQL were a language, SELECT would be your voice. It allows you to retrieve data from a database. Simple, yet powerful. 🔹 Basic syntax: SELECT column_name FROM table_name; 🔹 Example: SELECT name, sales_amount FROM orders; This means: ➡️ “Show me the name and sales amount from the orders table.” 💡 Why SELECT matters: - It’s the foundation of data analysis - Every insight starts with retrieving the right data - It’s used in almost every SQL query (yes, almost ALL). #SQL #LearningInPublic #100Daysofanalysis #DataAnalyst #SelectCommand

Your SQL query isn’t slow… it’s just doing too much work. Most performance issues don’t come from complex logic—they come from small, overlooked habits. This visual highlights 10 simple SQL optimization techniques that make a big difference:  🞄 Avoid SELECT * → fetch only what you need  🞄 Choose the right JOIN type → don’t over-fetch data  🞄 Limit results early (LIMIT / TOP)  🞄 Avoid unnecessary DISTINCT  🞄 Use EXISTS instead of COUNT  🞄 Optimize subqueries & derived tables  🞄 Index smartly (not blindly)  🞄 Avoid functions on indexed columns  🞄 Use UNION ALL instead of UNION 💡 Key Insight: SQL performance is less about rewriting queries… and more about reducing data movement and computation. 🔧 Practical takeaway: Think of your query like a pipeline:  🞄 Filter early  🞄 Reduce columns  🞄 Minimize joins  🞄 Let indexes do the work 📊 Example: Switching from SELECT * to specific columns + adding a proper index can drastically reduce execution time—especially in large datasets. Strong analysts don’t just get the right answer… they get it efficiently. #SQL #DataAnalytics #PerformanceTuning #DataEngineering #DatabaseOptimization #BigData #Analytics

This SQL question looks easy until you try to solve it I recently worked on an interesting problem: “Flag customers whose orders increase every month this year.” At first, it seems straightforward, but the real challenge lies in the logic. The goal is not just to find any increase, but to ensure consistent month over month growth. The key idea is to compare each month with the previous one and then verify that all comparisons show an increase. The core condition looks like this: HAVING COUNT(*) = COUNT(CASE WHEN cnt > prev_cnt THEN 1 END) This ensures that every month passes the increase check with no drops or exceptions. What I practiced through this: - Using DATE_TRUNC for monthly aggregation - Applying LAG for time based comparison - Writing conditional logic with HAVING - Focusing on logical correctness, not just working queries - You can find the full breakdown with sample data and outputs below I would be interested to know how others would approach this problem. #SQL #DataAnalytics #LearningInPublic #InterviewPrep

I wrote a SQL query to filter high-revenue countries… and it failed. The logic looked correct. But SQL threw an error. Here’s what I tried: 👉 Filtering total revenue using WHERE Something like: WHERE SUM(order_total) > 10000 And SQL didn’t accept it. That’s when I realized: 👉 I was filtering at the wrong stage of the query. In SQL, execution doesn’t happen the way we read the query. It actually works like this: FROM WHERE GROUP BY HAVING SELECT ORDER BY 💥 The mistake: WHERE runs before aggregation So it can’t use functions like SUM(), COUNT(), etc. ✅ The fix: Use HAVING for aggregated conditions: 👉 HAVING SUM(order_total) > 10000 💡 What I learned: WHERE filters rows HAVING filters grouped results Sounds simple… but easy to mess up in real queries. Now I think of it like this: 👉 WHERE → “filter raw data” 👉 HAVING → “filter summarized data” 📌 Lesson: If your query involves aggregation and filtering… Always ask: 👉 Am I filtering before grouping or after? This small distinction can save you from a lot of confusion. #SQL #DataEngineering #SQLTips #Analytics #LearnSQL #DataAnalytics #QueryOptimization #TechLearning #Debugging

This weekend, I’ll be revisiting one SQL topic that can be a little confusing at first but is very important to understand: JOINS In simple terms, “joins” help us combine data from different tables so we can get more meaningful insights. The common types I’m revising are: INNER JOIN – returns only the matching records from both tables LEFT JOIN – returns all records from the left table and the matching ones from the right RIGHT JOIN – returns all records from the right table and the matching ones from the left FULL JOIN – returns all matching and non-matching records from both tables CROSS JOIN – returns every possible combination of rows from both tables One thing I’m learning is that understanding joins is not just about memorising definitions. It’s about knowing when to use each one and what kind of result you want from your data. So this weekend is for more revision, more practice, and more clarity - one query at a time🤗 Which SQL concept are you currently revising or trying to understand better? #SQL #DataAnalytics #DataAnalysis #Omolabakethedataanalyst

SQL Execution Order (not how we write it, but how it actually runs) Most of us write queries like this: SELECT → FROM → WHERE → GROUP BY → ORDER BY But internally, SQL processes it very differently. SQL executes in this order: FROM JOIN WHERE GROUP BY HAVING SELECT DISTINCT ORDER BY LIMIT Here’s a simpler way to think about it FILTER → SHOW → SORT → LIMIT What this actually means • FILTER → FROM, JOIN, WHERE, GROUP BY, HAVING (Define data + reduce it step by step) • SHOW → SELECT, DISTINCT (Choose what you want to display) • SORT → ORDER BY (Organize the result) • LIMIT → LIMIT / TOP (Control how much data you return) Once we start thinking in execution order, we stop “trial and error” and start writing SQL with confidence. If you’re working with SQL daily, this mental model makes a huge difference. #SQL #DataAnalytics #LearnSQL #SQLTips #DataEngineering #Analytics

Maybe it’s time to stop using SQL indexes? Let’s first clarify what indexes are: Index — sorted version of a column in your table (obviously you don’t see it, but it’s there) 🗂 There are many ways you can make that sorted version: 1. Clustered Index  Created automatically based on a primary key 2. Non-clustered Index CREATE INDEX idx_users_last_name  ON users (last_name); That way every column we choose can be sorted under the hood. 3. Composite Index  Many columns and their order is very important. 4. Unique index Database will give you error if data is not unique 5. Full-Text index  Allows you to find word in the middle of the sentence. So what is the biggest problem with indexes? As I mentioned index is actually sorted copy of the column and every time you insert something, ITS DONE TWICE!  1 normal => 2 sorted column 👥 The basic principle with indexing: — getting data fast        BUT — inserting slow It’s always a trade off. Do you have any experience with spotting some useless indexes, or maybe finding them essential? #SQL #SoftwareEngineering #CodingSkills

Most analysts use SQL to pull data. The best analysts use SQL to answer business questions. Here's the difference: ❌ SELECT * FROM orders WHERE status = 'delayed' ✅ Which supplier is responsible for 80% of our delays — and what is it costing us? The query is just the tool. The real skill is knowing what question to ask before you open a database. A few principles I follow: → Start with the business outcome, not the table structure → Use window functions (RANK, LAG, PARTITION BY) to uncover trends — not just snapshots → Always validate your output against a known benchmark before presenting to stakeholders → Write CTEs over nested subqueries — your future self (and your team) will thank you SQL is not just a technical skill. It is a communication tool between raw data and business decisions. What is one SQL practice that has improved the quality of your analysis? I would love to hear in the comments. #BusinessAnalytics #SQL #DataAnalytics #MBALife #DataDriven

🚀 INNER JOIN in SQL If you're learning SQL, this is one of the most important concepts you must master. 🔍 INNER JOIN 👉 Returns only the rows that have matching values in both tables. 🔑 Key Points: 🔹 Matches rows based on a common column 🔹 Excludes non-matching records 🔹 Works like the intersection of two tables 🔹 Can join multiple tables together 🔹 Combines related data from multiple tables 💡 Example: 👉 The customer table has customer details. 👉 The orders table has order information. 🔹 But not all customers have orders 🔹 Not all orders have matching customers. 💻INNER JOIN Result Only records where Customer_id exists in both tables. Query: SELECT c.Customer_id, c.First_Name, o.Order_id, o.Amount FROM Customers c INNER JOIN Orders o ON c.Customer_id = o.Customer; Tips: INNER JOIN =Only matching data from both tables. #SQL #DataScience #DataEngineer #LearningSQL #TechTips #InterviewPrep #DataAnalytics