chore: Improve microbenchmark for string expressions

[andygrove]

Which issue does this PR close?

N/A

Rationale for this change

Make benchmarks faster to run and cover all string functions.

What changes are included in this PR?

Create the input data once instead of once per expression.

How are these changes tested?

Current benchmark results:

string.txt

[codecov-commenter]

Codecov Report

✅ All modified and coverable lines are covered by tests.
✅ Project coverage is 59.62%. Comparing base (f09f8af) to head (a0a220e).
⚠️ Report is 797 commits behind head on main.

Additional details and impacted files

@@             Coverage Diff              @@
##               main    #2964      +/-   ##
============================================
+ Coverage     56.12%   59.62%   +3.49%     
- Complexity      976     1375     +399     
============================================
  Files           119      167      +48     
  Lines         11743    15488    +3745     
  Branches       2251     2567     +316     
============================================
+ Hits           6591     9234    +2643     
- Misses         4012     4956     +944     
- Partials       1140     1298     +158     

☔ View full report in Codecov by Sentry.
📢 Have feedback on the report? Share it here.

🚀 New features to boost your workflow:

• ❄️ Test Analytics: Detect flaky tests, report on failures, and find test suite problems.