Resolution: No Action
Affects Version/s: V4.01_CSD02
Fix Version/s: None
Clarify that # characters in search phrases have to be percent-encoded because the first clear-text occurrence of the hash character (#) ends the query part of a URL and starts the fragment part, see https://tools.ietf.org/html/rfc3986#section-3.5.
The hash character (#) has meaning in both a URL and in OData so it makes sense that the hash character is not allowed in a search word. I would have expected though that the hash character could be used in a quoted search phrase, but the ABNF definition for search phrase does not include the hash character.
searchPhrase = quotation-mark 1*qchar-no-AMP-DQUOTE quotation-mark
qchar-no-AMP-DQUOTE = qchar-unescaped
/ escape ( escape / quotation-mark )
qchar-unescaped = unreserved / pct-encoded-unescaped / other-delims / ":" / "@" / "/" / "?" / "$" / "'" / "="
unreserved = ALPHA / DIGIT / "-" / "." / "_" / "~"
other-delims = "!" / "(" / ")" / "*" / "+" / "," / ";"
None of those definitions include the hash character. Was the hash character left out on purpose? Is the only way to include it to percent encode it?