google / google/google-sql-syntax-ts
Formatting of CTEs
- Dominant language
- TypeScript
- Stars
- 10
- Forks
- 4
- PR merge metrics
- No merged PRs in 30d
Description
Hello,
We use Dataform and BigQuery extensively in our projects. The built-in SQL formatter, which appears to be based on this repository, presents significant challenges in its current form due to how it handles Common Table Expressions (CTEs).
The primary issue is the lack of clear indentation, making it difficult to discern where CTEs begin and end in the formatted SQL. Below are examples that illustrate this problem:
### Example 1: BigQuery/Dataform Formatter
*The image shows a query formatted by the current formatter. Notice the minimal indentation, making it hard to distinguish the boundaries of the CTEs.*

### Example 2: SQLfluff Formatter
*The same query formatted with SQLfluff. The CTEs are clearly indented, making it easy to identify where they start and end.*

### Example 3: JetBrains Default Formatter (PyCharm/DataGrip)
*Another example using JetBrains' default formatter. It also clearly shows the structure of CTEs through proper indentation.*

### Suggestion
To enhance readability and usability, we suggest the following improvements for CTE formatting:
- **Indentation:** Clearly indent each level of CTEs.
- **Line Breaks:** Ensure that each CTE and its constituent SQL statements are well-separated by line breaks.
Implementing these changes would significantly improve the readability of formatted SQL queries, making it easier for developers to work with complex queries involving multiple CTEs.
Thank you for considering this suggestion.
Best regards,
Schnurres
Contributor guide
Research direction
The issue provides image examples of poorly indented CTEs and comparisons with SQLfluff and JetBrains formatters, but names no repository files or tests. Start by locating the SQL formatter's handling of Common Table Expressions and reproduce the reported BigQuery/Dataform case; done means nested CTEs have clear indentation and line breaks, with regression coverage for the formatting behavior.
Written by the indexing model from the issue text.
Assessment
- Tech stack
- sql, typescript
- Domain
- databases, tooling
- Issue type
- Feature
- Difficulty
- 4/5
- Estimated time
- 3-5 days
- Activity status
- Stale
- Clarity
- Mostly clear
- Newbie friendliness
- 35/100