Scrapes have different headers with same search parameters

Please see attached exported scrapes, run within 10 minutes of each other. Notice that one has the "groupType" header and the other doesn't. Why is this?

IT appears the "groupType" is a "for rent" thing as its an apartmetn complex and has a for rent listing. Why is this? Can we get rid of this option? I am running "sold" scrapes right now and don't want for rent listings. Why did that slip in there?

I am experiencing the same issue. This does not allow for standard import mapping to a database as the headers can change at random.

Hello, sorry for the inconvenience. Please add dev_no_strip=1 to your input. example: { "location": "New York", "limit": 10, "dev_no_strip": 1 }

Explanation: This actor normally will remove any empty values (NULL, FALSE, empty array/object and empty string) from the results. This is done to save space and memory. But this will make number of columns inconsistent, from one run to another. dev_no_strip will disable this behavior, and will keep empty values in datasets, so number of columns will consistent from 1 run to another. The dev_no_strip flag is hidden parameter and there is no UI for that (yet). I will soon update the actor UI to include this.

The "stripping" process is done before the results sent to Dataset Storage.

I hope this make sense. :)

Another way is to "re-shape" the dataset using dev_transform_enable and dev_transform_fields

@cat(Jupri) thanks for the suggestions. I'm definitely going to try the custom fields...would save a little time shaping the output file.