WebKoding

A different angle from the export-and-reimport route, and one that answers your "must not change content, URLs, SEO settings or categories" requirement directly: WP-CLI can write post meta without ever touching the post row.

First, find the real meta keys. Open one recipe that already has the five fields filled in and run:

wp post meta list

Recipe themes and plugins prefix their keys, so the field labelled "Preparation Time" in the editor is very often not stored as prep_time. This matters more than the rest of the job: if you guess the key, the write will succeed, the values will be in the database, and the template will still show nothing - and you will not discover it until all 170 are done.

Then export the list to review in a spreadsheet:

wp post list --post_type=post --format=csv --fields=ID,post_title > recipes.csv

Two things worth deciding in the spreadsheet rather than in WordPress. Total time is preparation plus cooking, so calculate that column instead of asking anything to estimate it - one less column to review and one less thing to get wrong. And since you are filling the sheet by hand, "only update the empty ones" is decided there: leave a row blank and no command is generated for it. That is a simpler guarantee than configuring an importer to skip non-empty fields.

Then write them one field at a time:

wp post meta update ""

That command writes to the postmeta table and nothing else. It cannot change post_content, the slug, the categories, the featured image, or the fields an SEO plugin keeps in its own meta - which is the assurance an import cannot give you as cheaply, because an importer is capable of touching the post and you are relying on having configured it not to.

Do five posts first and look at them on the front end before running the rest. If the values are in the database but the template is still empty, the key is wrong, and you have five to fix rather than 170.