wbeem

The issue involved blank product entries appearing in the WooCommerce → Reports → Stock → Out of Stock view. Upon investigation, I identified that these entries were linked to orphaned product variation records in the database — specifically, product_variation posts that had no post_title and were not associated with valid parent products.

After backing up the database, I safely deleted 116 such orphaned rows directly via phpMyAdmin. This resolved the issue — all blank rows are now gone from the Out of Stock report, and only valid product titles are displayed.

No core WooCommerce files were altered, and the database is now clean.

I’m sorry for the delayed response. Generating a new test site was very problematic. I have a temporary site on the Updraft Plus servers, but that will expire soon.

The displays for the Analytics Stock reports are VERY different now. The production site shows NO DATA TO DISPLAY. The same report on the cloned site 38 pages with 948 products. It’s the same data underneath, but with different displays. I cleared the Analytics Cache on the PROD server, but it didn’t change.

The files are both roughly the same size, but there are some variations. It turns out that’s because of the image URL differences between servers. Here’s part of the report I got from Grok:

The text content of the Description (excluding URLs) remains identical. The only difference is the domain name in the hyperlinks, with “PROD” using the production site (doogieshats.com) and “TEST” using a test site (vgn186j845n.c.updraftclone.com).

Summary of Differences

Based on the provided sample data:

• Rows: No evidence of rows present in one file but not the other.
• Columns:
  • Description: URLs differ, with “PROD” using https://doogieshats.com or http://doogieshats.com and “TEST” using https://vgn186j845n.c.updraftclone.com. Other text is identical.
  • Images: URLs differ, with “PROD” using https://DoogiesHats.com and “TEST” using https://vgn186j845n.c.updraftclone.com. Image paths and filenames are the same.
  • Other Columns: No differences detected in the sample.

What shall we do next about this issue?

Thanks for those ideas. I’m working through these suggestions on our TEST clone of the production site.

All tables complete the Check with OK status in both TEST and PROD sites.

We do not seem to have a wp_woocommerce_stock_quantities on the TEST site or the PROD site. We have stock_logs, but nothing else with “stock” in the table.

I tried the command DELETE FROM wp_postmeta WHERE meta_value = ”; on the TEST site. No effect.

From the command: SELECT * FROM wp_postmeta WHERE meta_key = ‘_sku’ AND meta_value = ”;

On the TEST site, I had zero results. On the PROD site, the command returned 129 rows. That’s because I already ran the DELETE command on TEST. As I mentioned, it did not affect the stock report.

Let me know what you suggest next. Thanks.

Now how am I supposed to check an invisible product to see if it’s in stock or not? There are no empty records on display in the Product area. If I filter products by Out of Stock status, I have 81 records.

The CSV export I shared earlier showed 116 empty or blank records overall.

System Status for TEST site: https://pastebin.com/HadDfrY1

System Status for PROD site: https://pastebin.com/z2K89FBu

Product Export ZIP – https://www.dropbox.com/scl/fi/n6xlh09ig2c2fleiybt7m/Doogies-Hats-CSV.zip?rlkey=2beg06hp786sy8g2vvmfemf3o&dl=0

Thanks,

William

I cloned the site to another server to use that for testing rather than use the production server. Since I need to disable many plugins, this seemed better than disrupting the customer experience.

After cloning the server, there was a different result to the missing entries on the production server and the testing server. I’ll share some screenshots.

This first link is the Out of Stock view from the Production server:

https://www.dropbox.com/scl/fi/mhe78y0tzn7y9lda9enyu/Screenshot-2025-04-04-at-5.07.41-PM.png?rlkey=4odcg2a28ne3sf1henyd7r1p2&dl=0

This next view is the same Out of Stock view, but from the cloned server:

https://www.dropbox.com/scl/fi/fxf2gyl9qiixs0n40uro7/Screenshot-2025-04-04-at-5.09.13-PM.png?rlkey=ztqueelsodnyurqv7876k5tjr&dl=0

Notice many rows are empty on the Production server, but a lot of the data (not all) shows on the Testing server.

The production server is a dedicated hardware server with 8 CPUs and 4 GB RAM. The testing server is a VPS with 2 cores and 4 GB RAM. Resources do not seem to be the issue. On the Production server, we don’t see any resource utilization going above 18% of capacity for CPU or RAM.

I performed the Export on both PROD and TEST servers. They should be identical, but there are slightly more records in TEST than PROD. I had Claude AI analyze the two files and here is its response.

I’ll analyze both CSV files to check if they are identical in content. Let me first examine the files.

Based on my analysis of both CSV files, I can confirm that they are not identical. Here are the key differences:

1. Row Count Difference:
   • PROD file contains 939 rows
   • TEST file contains 945 rows
   • The TEST file has 6 more rows than the PROD file
2. Headers: The headers in both files are identical. All column names match between the two files.
3. Content Differences: Looking at the sample rows, I noticed that while most of the data appears to be the same between matching records, there are some URL differences:For example, in the product with ID 2693 (Classic Blue Bandana Welding Beanie Caps):
   • In PROD file: The shop link is to “http://DoogiesHats.com” and images link to “https://DoogiesHats.com/wp-content/uploads/…”
   • In TEST file: The shop link is to “https://doogie.suburbiapress.com” and images link to “https://doogie.suburbiapress.com/wp-content/uploads/…”

This suggests that the TEST file may be using a different domain (suburbiapress.com) compared to the PROD file (DoogiesHats.com).

Summary of Differences:

1. TEST file contains 6 more product records than PROD
2. Domain/URL differences between the files (DoogiesHats.com vs. doogie.suburbiapress.com)
3. The structure and most content appears to be the same between matching records

I asked Claude to analyze the 6 new records in TEST. Turns out there were 8 new records, but there hasn’t been any data entry to the TEST server. It’s possible the production server deleted some records, but I’ll need to confirm if that happened. Here’s the result:

I’ll analyze the CSV files further to identify the specific 6 records that are present in the TEST file but missing from the PROD file.

I’ve identified the differences between the two CSV files. There are actually 8 extra records in the TEST file and 2 records in the PROD file that are not in TEST.Records in TEST but not in PROD (8 records):

These are all new welding cap products, consisting of 4 pairs of variable products and their variations:

1. Southwestern Sunset Band Welders Cap
   • ID: 366496 (variable product)
     • SKU: 333510
     • Categories: All Hats, NEW HATS, Standard Print Fabrics
   • ID: 366497 (variation)
     • Regular price: $21.20
     • Parent: 333510
2. Southwest Desert Dreams Band Welders Cap
   • ID: 366499 (variable product)
     • SKU: 333511
     • Categories: All Hats, NEW HATS, Standard Print Fabrics
   • ID: 366500 (variation)
     • Regular price: $21.20
     • Parent: 333511
3. Southeastern Feathers Orange Band Welders Cap
   • ID: 366502 (variable product)
     • SKU: 333512
     • Categories: All Hats, NEW HATS, Standard Print Fabrics
   • ID: 366503 (variation)
     • Regular price: $21.20
     • Parent: 333512
4. Southwestern Feathers Turquoise Band Welders Cap
   • ID: 366507 (variable product)
     • SKU: 333513
     • Categories: All Hats, NEW HATS, Standard Print Fabrics
   • ID: 366508 (variation)
     • Regular price: $21.20
     • Parent: 333513

Records in PROD but not in TEST (2 records):

• ID: 378850
• ID: 378853

All the new products in TEST appear to be part of a new product line, as they’re categorized under “NEW HATS” in addition to “All Hats” and “Standard Print Fabrics.” Each variable product has a corresponding variation record, and all are priced at $21.20.

The pattern suggests these are new products that have been added to the TEST environment but have not yet been deployed to PROD. Meanwhile, there are 2 records in PROD that aren’t in TEST, which could indicate either products that were removed from TEST before being added to PROD, or recent additions to PROD that haven’t been synchronized back to TEST.

Here’s the analysis of the 2 records in PROD, but not in TEST. They seem to be new records added recently:

I’ll analyze the two records that are in the PROD file but missing from the TEST file.

Based on my analysis, I’ve identified the two products that are in the PROD file but missing from the TEST file. Both appear to be newly added welding caps:Records in PROD but not in TEST:

1. Motorcycles Welding Hat
   • ID: 378850
   • Type: simple product (not a variable product with variations)
   • SKU: 3337834
   • Regular Price: $24.70
   • Categories: All Hats, NEW HATS, Premium Fabrics
   • Tags: “Welders hat pipe-fitter biker 100% cotton, welders hats”
   • Image URL: https://DoogiesHats.com/wp-content/uploads/2025/02/motorcycle.png
   • Stock: 12 units
   • Published: Yes
2. Steel Balls Welding Cap
   • ID: 378853
   • Type: simple product
   • SKU: 333785
   • Regular Price: $21.20
   • Categories: All Hats, NEW HATS, Standard Print Fabrics
   • Tags: “Welders hat pipe-fitter biker 100% cotton, welders hats”
   • Image URL: https://DoogiesHats.com/wp-content/uploads/2025/04/steel-balls.jpg
   • Stock: 12 units
   • Published: Yes

Key Observations:

1. Both products were added to PROD but not TEST, which is unusual for a typical development workflow where products should first appear in TEST before PROD.
2. Both products have image paths that indicate they were uploaded in 2025 (February and April), suggesting they are very recent additions:
   • “/2025/02/motorcycle.png”
   • “/2025/04/steel-balls.jpg”
3. Both products are simple (not variable) products, unlike the 8 products found in TEST but not in PROD, which were pairs of variable products and variations.
4. The “Steel Balls Welding Cap” appears to be the most recent addition based on its image path (April 2025).
5. Both products are categorized under “NEW HATS” just like the products found in TEST but not in PROD.

This suggests that new products are being added to both environments independently rather than following a TEST-then-PROD workflow. The Steel Balls Welding Cap appears to be a particularly recent addition to PROD, given its April 2025 image date.

Overall, this does not explain why we see more records visible in our TEST server compared to our PROD server. The variance is more than 8 records.

I also asked Claude to compare the columns with an empty Name field. Seems to be identical:

I’ll analyze both CSV files to count how many rows have no entry in the Name column.

Based on my analysis of both CSV files:

PROD file:

• 116 rows out of 939 have no Name entry (12.35% of total rows)

TEST file:

• 116 rows out of 945 have no Name entry (12.28% of total rows)

All rows with empty Name fields are of Type “variation” in both files. Looking at the samples, these rows also have null SKU values, suggesting they are product variations that inherit key information from their parent products.

This is consistent with the expected structure for WooCommerce product exports, where variation products often don’t have their own names but instead inherit the name from their parent variable product, with attributes determining the specific variation (such as size, color, etc.).

The count of rows without Names is exactly the same (116) in both environments, despite the TEST file having 6 more rows overall. This indicates that the additional products in the TEST environment all have values in their Name fields, and none of the variation products added to TEST lack names.

The consistency in this pattern between environments suggests this is intentional data structure rather than missing data.

• This reply was modified 1 year, 1 month ago by wbeem.

Oh, thank you. This is my first time with OpenLiteSpeed and I wasn’t aware. Makes me wonder why some tools want to update the .htaccess file when they detected the change from Apache.

I’ll review the info you sent and try it out. I appreciate your help.

William