[fix](doc) zh: correct expected values in 5 function pages where docs drift from cluster output

[boluor]

Summary

Five ZH-only doc pages on `version-4.x` have a result block whose values disagree with what a single-node Apache Doris 4.1.1 cluster actually returns. In every case the EN counterpart already shows the correct value, so this PR is a pure ZH-side correction; the EN pages are untouched.

Page	Was (ZH)	Cluster on 4.1.1
`encrypt-digest-functions/murmur-hash3-64-v2.md`	column header `murmur_hash3_64(...)` + hash values from an earlier implementation	header is `murmur_hash3_64_v2(...)` and the values are `-3758069500696749310` and `-662943091231200135`
`string-functions/instr.md`	`INSTR('ṭṛì ḍḍumai hello', 'hello')` claimed `13`	`12` (the `h` is the 12th character)
`string-functions/length.md`	`LENGTH('ṭṛì')` claimed `9`	`8` (UTF-8 bytes: ṭ=3 + ṛ=3 + ì=2)
`string-functions/locate.md`	`LOCATE('o', 'foobar', 2)` claimed `3`	`2` (the `o` at position 2 satisfies "starting at or after position 2")
`string-functions/lpad.md`	`LPAD('X', 7, 'HELLO')` claimed `HELLOX` (length 6)	`HELLOHX` (length 7)

Verification

Every snippet was run against a single-node Apache Doris 4.1.1 cluster — each updated result block is the verbatim mysql client output.

Test plan

• Each ZH snippet on its updated page is run on 4.1.1 — result matches.
• No EN-side change.
• Each file is a 1- or 2-line edit (or, for murmur-hash3, a header+row swap); no surrounding prose touched.

🤖 Generated with Claude Code