mirror of
https://github.com/meilisearch/MeiliSearch
synced 2024-11-05 12:38:55 +01:00
Merge #637
637: We avoid skipping errors in the indexing pipeline r=ManyTheFish a=Kerollmops This PR is related to https://github.com/meilisearch/meilisearch/issues/2764 and should fix it when merged into Meilisearch. Co-authored-by: Kerollmops <clement@meilisearch.com>
This commit is contained in:
commit
2865b063ad
@ -304,24 +304,23 @@ where
|
||||
documents_chunk_size,
|
||||
);
|
||||
|
||||
let result = original_chunk_iter
|
||||
.and_then(|original_chunk_iter| Ok((original_chunk_iter, flattened_chunk_iter?)))
|
||||
.map(|(original_chunk, flattened_chunk)| {
|
||||
// extract all databases from the chunked obkv douments
|
||||
extract::data_from_obkv_documents(
|
||||
original_chunk,
|
||||
flattened_chunk,
|
||||
pool_params,
|
||||
lmdb_writer_sx.clone(),
|
||||
searchable_fields,
|
||||
faceted_fields,
|
||||
primary_key_id,
|
||||
geo_fields_ids,
|
||||
stop_words,
|
||||
max_positions_per_attributes,
|
||||
exact_attributes,
|
||||
)
|
||||
});
|
||||
let result = original_chunk_iter.and_then(|original_chunk| {
|
||||
let flattened_chunk = flattened_chunk_iter?;
|
||||
// extract all databases from the chunked obkv douments
|
||||
extract::data_from_obkv_documents(
|
||||
original_chunk,
|
||||
flattened_chunk,
|
||||
pool_params,
|
||||
lmdb_writer_sx.clone(),
|
||||
searchable_fields,
|
||||
faceted_fields,
|
||||
primary_key_id,
|
||||
geo_fields_ids,
|
||||
stop_words,
|
||||
max_positions_per_attributes,
|
||||
exact_attributes,
|
||||
)
|
||||
});
|
||||
|
||||
if let Err(e) = result {
|
||||
let _ = lmdb_writer_sx.send(Err(e));
|
||||
|
Loading…
Reference in New Issue
Block a user