Problem counting duplicate fields

[Marcel1]

Hi

I have a database of gene sequences, and it would be useful for me to be able to identify identical sequence strings. Presently, each record in the table "sequences" contains a gene sequence as a text variable, and additional fields with information about the sequence.

I thought I had solved my problem by creating a second occurance of the table Sequences with a self-join to the sequence field, then counting the number of matches to this second table with the calculated field "duplicate sequences", defined as "Count(SequencesTable2::Sequence) - 1"

However, I am finding that FM only seems to look at the first 109 characters of this field to decide if there is a match. So, sequences that are identical within the first 109 characters, but differ further on in the string are still counted as duplicates. Is there any way around this problem?

Many thanks,

Marcel

[SlimJim]

There is a limit as you have found on how far FMP indexes field content. In your particular circumstance there is a possible solution but it may not be feasible. I have no idea how long a gene sequence is which is why I don't know if this is feasible. My suggestion is to split the sequences into strings of length 100 (say) and put these into a sequence of fields then use the complete set of fields in the self join.

Two sequences will be the same if and only if the sequence of fields containing the length-100 strings match in order. If this is feasible then you have a self-join via duplicate gene sequences.

One snag that might occur is if the sequence lengths differ by enough so that some of the subsequence fields are empty. FMP does not match empty fields so that would break the relationship.

[Marcel1]

Thanks, I will give this a try.

Marcel