What it does
Reads a table of an export made with Export for AI or with the "Export for AI" window, without loading it whole: it scans the CSV file and returns only the rows needed. It can keep only the rows with given values in given columns (equality, case-insensitive), those that contain a text in any cell, only some columns, and page through the result (200 rows at a time unless told otherwise).
With table "indice" it returns the text of 00_INDEX.md, which describes all the tables and columns of the export. If no folder is given, it reads the most recent export (with an index) in the exports folder.
It returns the subfolder read, the table file, the columns, how many rows match in total, the row the page starts from and the rows of the page.
How it works
- Your AI reads the index first, then picks the table and the columns or filters that fit your question.
- BIMit opens the CSV file on disk and returns the matching rows; Revit and the model are not queried.
- You get your AI's answer; it can fetch more pages when more rows match than were returned.
Example requests
- "In the last export, which walls are on Level 1? Give me id, type and length."
- "Search the warnings for those about overlapping walls."
- "Which views are not on any sheet?"
What changes and how to undo
Nothing: the tool only reads the export files. The model and the files stay as they are.
Limits
AI level required: Read.
- It reads only exports that exist on disk: if there are none, export first.
- If the table was not included in that export, or a column does not exist, BIMit says so and lists the available tables or columns.
- The data are those of the moment of the export: changes made to the model afterwards are not there.
Parameters
| Parameter | Meaning |
|---|---|
tabella | Required. File or key of the table (e.g. 06_elements, 07_element_parameters.csv), or "indice" (the index). |
cartella | The export subfolder (returned by Export for AI). Default: the most recent one in the exports folder. |
filtri | Only the rows with these values: {column: value} (equal, case-insensitive), e.g. {"category": "Walls"}. |
cerca | Only the rows that contain this text in any cell. |
colonne | Only these columns. |
da | Skip the first N matching rows (paging). |
righe | How many rows to return (default 200). |