A common thing with Bigtable is the ability to scan based on a prefix, which is basically a range with start=prefix and end=someAlteration(prefix), but its tricky to do manually.
Any chance we can add a prefix filter that "just works"?
Go does this a bit trickily (see https://github.com/GoogleCloudPlatform/google-cloud-go/blob/master/bigtable/bigtable.go#L307), the gist seems to be (for those not familiar with go):
- if the string is empty, the end range marker should just be empty string
- otherwise, find the last character in the string that can be incremented (ie,
abcde can increment position 4, abc\xff\xff can increment position 2.
- using that pivot, take
start.substr(0, n-1) and append the incremented character (start[n]+1)
I think -- in JS -- this would look something like....
var getPrefixEndRange = function(start) {
var maxChar = String.fromCharCode(0xff);
var position = start.length-1;
// Walk backwards until we get to a character we can increment.
while (start[position] == maxChar && position >= 0) position--;
// If the position is -1, there is no reasonable end range for the prefix.
if (position == -1) return '';
var nextChar = String.fromCharCode(start.charCodeAt(position)+1)
return start.substring(0, position) + nextChar;
}Some test cases...
getPrefixEndRange('start'); // -> 'staru'
getPrefixEndRange('X' + String.fromCharCode(0xff)); // -> 'Y'
getPrefixEndRange('xoo' + String.fromCharCode(0xff)); // -> 'xop'
getPrefixEndRange('com.google.'); // -> 'com.google/'
getPrefixEndRange(String.fromCharCode(0xff)); // -> ''
getPrefixEndRange(''); // -> ''
A common thing with Bigtable is the ability to scan based on a prefix, which is basically a range with start=prefix and end=someAlteration(prefix), but its tricky to do manually.
Any chance we can add a
prefixfilter that "just works"?Go does this a bit trickily (see https://github.com/GoogleCloudPlatform/google-cloud-go/blob/master/bigtable/bigtable.go#L307), the gist seems to be (for those not familiar with go):
abcdecan increment position 4,abc\xff\xffcan increment position 2.start.substr(0, n-1)and append the incremented character (start[n]+1)I think -- in JS -- this would look something like....
Some test cases...