Cloud Bigtable: CheckAndMutateRow : ExactValueFilter

[sumith-tempus]

Steps to reproduce

1. Run the given code example to re-produce the error

ExactValueFilter is a subclass of ValueRegexFilter but I think I believe it should not care about when the filter is ExactValueFilter. I am not quite confident whether this is an issue or a feature request.

I am trying to understand, Is there a way to do CheckAndMutateRow for the column value(update if and only if the value is changed)

Code example

client = bigtable.Client()
instance = client.instance("instance-name")
table_id = "table-name"
table = instance.table(table_id)
row_key = "row_key_1"
col_name = "col_name1"
cell_val = "test(value.pdf"
col1_filter = row_filters.ColumnQualifierRegexFilter(col_name)
label1_filter = row_filters.ExactValueFilter(cell_val)
chain1 = row_filters.RowFilterChain(filters=[col1_filter, label1_filter])

row_obj = table.conditional_row(row_key, filter_=chain1)
row_obj.set_cell("test_family_1", col_name, cell_val, state=False)
row_obj.commit()

Stack trace

/var/folders/t7/znmnfd3x2191zm8hybnnt5fx8zypkn/T/ipykernel_49224/1449758387.py in <module>
----> 1 row_obj.commit()

/usr/local/lib/python3.9/site-packages/google/cloud/bigtable/row.py in commit(self)
 590 
 591 data_client = self._table._instance._client.table_data_client
--> 592 resp = data_client.check_and_mutate_row(
 593 table_name=self._table.name,
 594 row_key=self._row_key,

/usr/local/lib/python3.9/site-packages/google/cloud/bigtable_v2/services/bigtable/client.py in check_and_mutate_row(self, request, table_name, row_key, predicate_filter, true_mutations, false_mutations, app_profile_id, retry, timeout, metadata)
 999 
 1000 # Send the request.
-> 1001 response = rpc(request, retry=retry, timeout=timeout, metadata=metadata,)
 1002 
 1003 # Done; return the response.

/usr/local/lib/python3.9/site-packages/google/api_core/gapic_v1/method.py in __call__(self, *args, **kwargs)
 143 kwargs["metadata"] = metadata
 144 
--> 145 return wrapped_func(*args, **kwargs)
 146 
 147 

/usr/local/lib/python3.9/site-packages/google/api_core/grpc_helpers.py in error_remapped_callable(*args, **kwargs)
 67 return callable_(*args, **kwargs)
 68 except grpc.RpcError as exc:
---> 69 six.raise_from(exceptions.from_grpc_error(exc), exc)
 70 
 71 return error_remapped_callable

/usr/local/lib/python3.9/site-packages/six.py in raise_from(value, from_value)

InvalidArgument: 400 Error in chained row filter googleapis/python-bigtable#1 : Error in field 'value_regex_filter' : Invalid RE2 regular expression 'test(value.pdf': missing ): test(value.pdf

Happy to Provide more information if needed

[Mariatta]

This might have been due to poor documentation about the ExactValueFilter. As you found out, indeed it derives from the ValueRegexFilter, so it expects a valid regex still. You can escape the ( in the value:

cell_val = "test\(value.pdf"
label1_filter = row_filters.ExactValueFilter(cell_val)

should do it.
Alternatively, can you try using the ValueRangeFilter ?

cell_val = "test(value.pdf"
row_filters.ValueRangeFilter(start_value=cell_val, end_value=cell_val)

[igorbernstein2]

ExactValueFilter seems to have very broken semantics. I dont think we can fix it in place (in case other customers depend on the broken behavior). Instead we will provide a new filter with correct semantics:
googleapis/python-bigtable#620 (comment)

[parthea]

Transferring to google-cloud-python as the code for python-bigtable will be migrated there in 1-3 weeks.

[daniel-sanche]

IIRC, this was addressed by adding a LiteralValueFilter. But let me know if this is still an issue