1
$\begingroup$

I would like to run the association rule mining algorithm of the Orange library on a dataset that is stored in a PostgreSQL database. The table 'buildingset' contains the itemsets for each user, thus each record is related to a user, and each field is related to an item. The values are either 1 (smallint) or missing. The table has about 14,000 records and 31 fields.

When I try to run the algorithm on this dataset, I get the following error:

ValueError Traceback (most recent call last): File "/opt/orange/orange3/Orange/canvas/scheme/widgetsscheme.py", line 722, in process_signals_for_widget handler(*args) File "/home/bdukai/.local/lib/python3.4/site-packages/orangecontrib/associate/widgets/owassociate.py", line 444, in set_data self.X = data.X File "/opt/orange/orange3/Orange/data/sql/table.py", line 353, in X self.download_data(AUTO_DL_LIMIT) File "/opt/orange/orange3/Orange/data/sql/table.py", line 333, in download_data raise ValueError("Too many rows to download the data into memory.") ValueError: Too many rows to download the data into memory.

Thus is there any way overcome this limitation, without upgrading the hardware?

$\endgroup$
  • $\begingroup$ could you share how you could connect with SQL table, I downloaded freeDTS ( don't know how to install it) as well as running pip install pymssql in command line ( completed with no issue). But nothing change. I try to use file widget, but I could only get 60K row out of 1 Million. thanks $\endgroup$ – oveno Jun 5 '18 at 9:54
  • $\begingroup$ I haven't used Orange since then and I honestly don't remember, sorry $\endgroup$ – Balazs Dukai Jun 14 '18 at 12:48
0
$\begingroup$

You can use "download data to local memory" checkbox in SQL Table widget, it should allow you to work with up to 1000000 rows.

enter image description here

| improve this answer | |
$\endgroup$

Your Answer

By clicking “Post Your Answer”, you agree to our terms of service, privacy policy and cookie policy

Not the answer you're looking for? Browse other questions tagged or ask your own question.