[Feature Request]: Use thousands separators in input boxes where possible.

[usbrandon]

What would you like to happen?

This is a UX improvement request. Transforms like the Parquet Output require inputting large numbers. Would you please enhance the input boxes to format the input for the user using a thousand separators to make the numbers easier to verify?

Issue Priority

Priority: 3

Issue Component

Component: Hop Gui

[mattcasters]

I'm sorry Brandon. I'm at a loss about what exactly you're asking for here. Which large numbers do you need to enter in the Parquet Output Dialog? Things like the Row Group Size, and it's hard for you to see the exact size?
Perhaps we could add support for k,M,G suffixes?

[mattcasters]

Implement an alterative to Const.toInt() and Const.toLong() that replaces comma and decimal values. It also replaces trailing k with 000, m with 000000 and g with 000000000, or even x10^8 etc.

[mattcasters]

[mattcasters]

Expanded integer notation (implementation overview)

Addressing the UX pain of typing large counts (e.g. one hundred million) in dialogs like Generate Rows and Parquet Output.

Format support

New opt-in parsers in Const (existing toInt / toLong stay strict):

Form	Examples	Result
Grouping	100,000,000, 100.000.000, 100_000_000, 100 000 000	100000000
Suffixes	100k, 100m, 1g, 1b	×10³ / 10⁶ / 10⁹
Fractional + suffix	1.5m, 1,5m	1500000
Scientific	1e8, 1E8, 1x10^8, 10^8	100000000

API: Const.expandIntegerString(), toLongExpanded(), toIntExpanded().

Variables (${...}) are still resolved first, then expanded.

GUI indicator

On TextVar fields that accept this notation, a small # icon appears next to the existing $ (variable) marker. Hovering # shows a tooltip with the allowed formats.

Where it is enabled (runtime + dialog where applicable)

Row / limit counts

• Generate Rows (row limit; also Beam Generate Rows handler)
• Table Input
• Clone Rows
• Cube Input, XML Input Stream (limit + rows to skip), SAS Input, Avro File Input, Salesforce Input
• File Metadata
• Dialogs that parse a row limit on OK: Text File Input, Excel Input, JSON Input / Normalize, YAML Input, Property Input, Get File Names, Get Subfolders, Load File Input, Tika, LDAP Input, Dynamic SQL Row, Database Join, Get XML Data, Google Analytics

Sizes / buffers / splits

• Sort Rows (sort size)
• Parquet Output (file split size, row group / data / dictionary page sizes)
• CSV Input (buffer size)
• MySQL / MonetDB / Snowflake / CrateDB / Oracle bulk loaders (bulk/buffer/split/commit/bind/read/max-errors as applicable)

Commit / batch / group sizes

• Table Output, Insert/Update, Update, Delete, Synchronize After Merge (commit size)
• Kafka Consumer (batch size)
• Neo4j Cypher / Cypher Builder / Graph Output / Neo4j Output (batch size)
• Azure Event Hubs listen/write (batch / prefetch)
• Reservoir Sampling (sample size)
• Pipeline Executor & Workflow Executor (group size)

Engines (config parse)

• Beam Flink max bundle size
• Beam Spark max records per batch / bundle size

Notes

• Opt-in only — ports, field lengths, timeouts, XML result counters, etc. still use strict toInt/toLong.
• Plain Text widgets (not TextVar) accept the notation on save where parsing was updated, but do not show the # badge until converted to TextVar.
• Branch: eil-friction on https://github.com/mattcasters/hop/tree/eil-friction (commit includes unit tests for Const + transform IT green).

Happy to open a PR against apache/hop main when ready.