Let's say I have a dataframe with column names as follows:

col_id_1, col_id_2, ..., col_id_m, property_1, property_2 ..., property_n

As an example, how would I search across all col_ids for, say, the value 5 (note that 5 won't appear in multiple col_ids in the same row), and then choose all rows that contain this value? On top of that, once I've found all rows that have a col_id containing the value 5, I'll combine all the col_ids with value 5 into a single id column, and also only choose, say, property_8 and property_25000 as additional columns.

In this case, I would have a table with the following columns:

id, property_8, property_25000

where the id column only contains rows with value 5. Is such a thing possible in pandas?

IIUC, first filter you columns by contain the col_id, then we using any check if any columns have the number 5

df.loc[df.filter(like='col_id').eq(5).any(1),['property_8','property_25000']].assign(id=5)