Implement column-type-aware CSV import: defer type conversion to INSERT based on table definitions

[Copilot]

Summary

Applied maintainer feedback to further optimize the type converter utility.

Changes Made

Type Converter Optimizations (src/15utility.js)

1. Simplified int regex (Line 1355)

   • Before: /^int|int$|^smallint$|^bigint$|^tinyint$/i
   • After: /int$/i
   • All int types (smallint, bigint, tinyint) are already covered by int$ pattern

2. Removed trim() call (Line 1376)

   • Before: var trimmed = value.trim(); if (trimmed == +trimmed)
   • After: if (value == +value)
   • Direct comparison without unnecessary trim operation

3. Boolean conversion using regex (Lines 1396-1397)

   • Before: var lower = value.toLowerCase(); return lower === 'true' || lower === '1' || lower === 'yes';
   • After: return /^(true|1|yes)$/i.test(value);
   • More efficient single regex test instead of toLowerCase + multiple comparisons

Testing Results

• ✅ All 17 CSV string type tests pass
• ✅ All 47 CSV-related tests pass
• ✅ No regressions detected
• ✅ Code is cleaner and more efficient

Original prompt

---

This section details on the original issue you should resolve

<issue_title>Wrong datatype after reading FROM CSV </issue_title>
<issue_description>I am very enthusiastic about alasql and it was already a great help to me. Thanks a lot for the great work.
But...

Datatype switches from string to number after FROM CSV import.

CSV sample:
"117.20";"some name"
"88.33";"other name"

SQL:
CREATE TABLE test (id string,name string);
SELECT * INTO test FROM CSV('some csv file');
SELECT * FROM test;

results in:

id	name
117.2	some name
88.33	other name

The column 'ID' datatype changed to number and the value is handled as float and trailing Null is cut off.
Bad if this is an identifier.

PS: it seems to work with FROM XLSX, when column is formated as TEXT in Excel
</issue_description>

Comments on the Issue (you are @copilot in this section)

@mathiasrw Hi @casavipiestrick - thank you for the kind words.

Good point. Nice little bug there :)</comment_new>
<comment_new>@mathiasrw
The conversion of numberstrings into numbers has been part of Alasql too long to change. I am happy to let a config flag change how it works.

@mathiasrw Any PR is welcome to fix this - but the fix must be provide in a way so you need to set something like alasql.config.voidstr2intCSV (preferably a better name) to have that new functionality.

• Fixes Wrong datatype after reading FROM CSV  #1181

---

💡 You can make Copilot smarter by setting up custom instructions, customizing its development environment and configuring Model Context Protocol (MCP) servers. Learn more Copilot coding agent tips in the docs.

[mathiasrw]

I recommend we keep csv data as string until we are inserting them.

If there are col definitions, then we adhire to that (also with things like dates, int, string, float, varchar)

If there are no definitions we behave like now when we do auto convert if we detect numbers (for example into tables provideda s reference data)

[mathiasrw]

@copilot

[Copilot]

I recommend we keep csv data as string until we are inserting them.

If there are col definitions, then we adhire to that (also with things like dates, int, string, float, varchar)
...

Implemented as requested. CSV data now stays as strings until INSERT. Type conversion happens based on column definitions (STRING, INT, FLOAT, etc.) when available. Without definitions, the csvStringToNumber option controls auto-conversion for backward compatibility. Commit: 588303a

[mathiasrw]

Small int, big int, tinyint are all coverd by int$

[Copilot]

Simplified int regex to /int$/i which covers all int types. Commit: 689a7ca

[mathiasrw]

Please avoid trim here.

[Copilot]

Removed trim() call, now using direct comparison value == +value. Commit: 689a7ca

[mathiasrw]

Can we please use a regex here instead?

[Copilot]

Replaced toLowerCase + string comparisons with single regex: /^(true|1|yes)$/i.test(value). Commit: 689a7ca