Function Don't Work When Applied To Rows, Only Result For First Row Is Returned

Mar 9, 2025 by ADMIN 80 views

**Function Doesn't Work When Applied to Rows: A Bug in Netquack Extension**

Introduction

The Netquack extension in DuckDB is a powerful tool for extracting information from strings. However, a bug has been discovered where the macro functions only work on scalar values and not on rows. When applied to rows, only the result for the first row is returned, followed by a segfault. In this article, we will delve into the details of this bug, provide steps to reproduce it, and discuss the expected behavior.

Describe the Bug

The macro functions in Netquack only seem to work on scalar values and not on rows. When applied to rows, only the result for the first row is returned. This is typically followed by a segfault. This behavior is not expected, as macro functions should return the result for every row.

To Reproduce

The steps to reproduce this behavior are as follows:

Step 1: Create a Simple Table of URLs

Create a simple table of URLs using the following SQL query:

CREATE OR REPLACE TABLE urls AS
SELECT 'a.example.com' AS url
UNION ALL
SELECT 'https://b.a.example.com/path/path';

This will create a table with two rows, each containing a URL.

Step 2: Query the Table

Query the table using the following SQL query:

SELECT extract_host(url), url
FROM urls;

This will return the host extracted from each URL, along with the original URL.

Expected Result

The expected result is:

┌───────────────────┬───────────────────────────────────┐
│ extract_host(url) │                url                │
│      varchar      │              varchar              │
├───────────────────┼───────────────────────────────────┤
│ b.a.example.com   │ https://b.a.example.com/path/path │
│ a.example.com     │ a.example.com                     │
└───────────────────┴───────────────────────────────────┘

However, the actual result is:

┌───────────────────┬───────────────────────────────────┐
│ extract_host(url) │                url                │
│      varchar      │              varchar              │
├───────────────────┼───────────────────────────────────┤
│ a.example.com     │ a.example.com                     │
│                   │ https://b.a.example.com/path/path │
└───────────────────┴───────────────────────────────────┘

As you can see, the host is only extracted for the first row.

Workaround

To get the expected result, you can use the following SQL query:

SELECT extract_host(url), url
FROM (SELECT url FROM urls ORDER BY url DESC) AS t;

This will return the host extracted from each URL, along with the original URL.

Expected Behavior

The expected behavior is that the macro functions should return the result for every row. In this case, the extract_host function should return the host extracted from each URL, along with the original URL.

Versions

The versions of Netquack and DuckDB used to reproduce this bug are:

Netquack Version: d7f4bec
DuckDB Version: v1.2.1

Additional Context

This bug was reproduced on an M1 Mac running macOS 15.3.

Conclusion

Introduction

In our previous article, we discussed a bug in the Netquack extension of DuckDB where the macro functions only work on scalar values and not on rows. When applied to rows, only the result for the first row is returned, followed by a segfault. In this article, we will answer some frequently asked questions about this bug.

Q: What is the Netquack extension?

A: The Netquack extension is a powerful tool for extracting information from strings in DuckDB. It provides a set of macro functions that can be used to extract information from strings, such as extracting hosts, domains, and paths.

Q: What is the bug in the Netquack extension?

A: The bug in the Netquack extension is that the macro functions only work on scalar values and not on rows. When applied to rows, only the result for the first row is returned, followed by a segfault.

Q: How do I reproduce the bug?

A: To reproduce the bug, you can follow these steps:

Create a simple table of URLs using the following SQL query:

CREATE OR REPLACE TABLE urls AS
SELECT 'a.example.com' AS url
UNION ALL
SELECT 'https://b.a.example.com/path/path';

Query the table using the following SQL query:

SELECT extract_host(url), url
FROM urls;

Observe that the host is only extracted for the first row.

Q: What is the expected behavior?

A: The expected behavior is that the macro functions should return the result for every row. In this case, the extract_host function should return the host extracted from each URL, along with the original URL.

Q: How do I get the expected result?

A: To get the expected result, you can use the following SQL query:

SELECT extract_host(url), url
FROM (SELECT url FROM urls ORDER BY url DESC) AS t;

This will return the host extracted from each URL, along with the original URL.

Q: What are the versions of Netquack and DuckDB used to reproduce this bug?

A: The versions of Netquack and DuckDB used to reproduce this bug are:

Netquack Version: d7f4bec
DuckDB Version: v1.2.1

Q: What is the additional context for this bug?

A: This bug was reproduced on an M1 Mac running macOS 15.3.

Q: Is there a workaround for this bug?

A: Yes, there is a workaround for this bug. You can use the SQL query mentioned above to get the expected result.

Q: Is this bug fixed in the latest version of Netquack?

A: We do not have information on whether this bug is fixed in the latest version of Netquack. However, we recommend checking the latest version of Netquack to see if the bug has been fixed.

Conclusion

In conclusion, the Netquack extension has a bug where the macro functions only work on scalar values and not on rows. When applied to rows, only the result for the first row is returned, followed by a segfault. We hope this Q&A article has provided you with the information you need to understand and work around this bug.