fix: no old_product_name_token found#7
Conversation
|
Hi, since we are using duckdb here, line 1191 will actually give you access to table old_product_name_token here, so this shouldn't be a problem |
|
@CurryTang Thank you for the quick reply! However, when running No table named Am I missing something? I am using dataset_path=data
export DBB_DATASET_HOME=$dataset_path/diginetica/raw
export DBB_PROJECT_HOME=$dataset_path
mkdir -p "$dataset_path/diginetica/raw"
mkdir -p "$dataset_path/diginetica/old"
mkdir -p "$dataset_path/diginetica/expert"
mkdir -p "$dataset_path/diginetica/old/data"
mkdir -p "$dataset_path/diginetica/expert/data"
mkdir -p "$dataset_path/diginetica/old/ctr"
mkdir -p "$dataset_path/diginetica/expert/ctr"
mkdir -p "$dataset_path/diginetica/old/purchase"
mkdir -p "$dataset_path/diginetica/expert/purchase"
python3 -u -m dbinfer.main download diginetica
python3 -u -m main.preprocessing_dataset diginetica |
|
Oh, I see the problem, we should use duckdb.sql instead of duckdb.query here. It should fix the problem and later we will release a new version duckdb.sql("XXX").df() |
|
@CurryTang Same error with new_product = duckdb.sql("""
SELECT np.itemId,
np.categoryId,
np.pricelog2,
array_agg(opt.token) AS name_tokens
FROM new_product AS np
LEFT JOIN old_product_name_token AS opt
ON np.itemId = opt.itemId
GROUP BY np.itemId, np.categoryId, np.pricelog2
""").df() |
|
That's weird. The final fix would be adding this variable to the context of duckdb. I will check this after a deadline. |

Error:
no old_product_name_token found
Description of changes:
add product_name_token from expert schema
Could you please confirm if this fix is correct?

If not, where should I look for the missing old_product_name_token in the query below?
Thanks a lot!