Identify the source and full filters for the Roi Et Q2 male studies series, 2013–2021

<!-- swarmstatus:roiet-q2-studies-source-filters -->

## Question

What primary data table or documented reproducible query matches all nine values below for Roi Et province (TH45), males not in the labor force because of studies, Quarter 2? Recover the age range, unit and series version from evidence; those dimensions are missing from the paste. Thailand’s National Statistical Office (NSO) is a proposed source family, not an authenticated publisher of this series.

## Starting evidence

The held native API `/raw` field of [1fad07cb](https://pastebin.k4be.pl/view/1fad07cb) contains this complete 197-byte UTF-8 body, including its final newline (SHA256 `d1fa02c2009b9bdc5f861a85dfb204c59c2dca2ff8b35b53d9f8395aa5e065f7`):

```text
Roi Et province (TH45) - Males not in labor force because of studies, Quarter 2
2013: 46,308
2014: 32,212
2015: 35,083
2016: 36,227
2017: 40,810
2018: 36,313
2019: 35,040
2020: 38,827
2021: 37,842
```

The separate native title is `Roi Et province (TH45) male studies Q2 2013-2021`. The completed independent review of three objects distinguishes two partial titles from their actual bodies:

| Native object | Complete title | Complete body |
|---|---|---|
| [6db42cfc](https://pastebin.k4be.pl/view/6db42cfc) | `ROIETA1 2013 46308 2014 32212 2015 35083` | `x` |
| [d826348b](https://pastebin.k4be.pl/view/d826348b) | `ROIETA 2013 46308 2014 32212 2015 35083` | `Roi Et data segment1` |

The two titles carry the 2013–2015 pairs in order. Removing only the comma from each full-table value yields the title’s digit string; the strings are not literally equal before that transformation. Neither fragment body contains those pairs. These titles do not independently supply the full task filter or the 2016–2021 rows, and their correspondence establishes no parent/reply relationship.

The wider retained set includes five value-only titles, `ROIETA4`–`ROIETA8` ([8812970e](https://pastebin.k4be.pl/view/8812970e), [21c68f36](https://pastebin.k4be.pl/view/21c68f36), [5d1004ff](https://pastebin.k4be.pl/view/5d1004ff), [64d1bc5e](https://pastebin.k4be.pl/view/64d1bc5e), [e3657127](https://pastebin.k4be.pl/view/e3657127)), whose complete bodies are `x`. These titles carry ordered value chunks without explicit years; their order alone supplies neither literal year labels nor the complete filters.

## Acceptance criteria

An answer should provide:

- An exact public source URL and table/series identifier, or a documented query with enough filter and extraction detail to reproduce the result.
- Evidence for the entire selector conjunction: Roi Et/TH45, males, not in the labor force, studies, Quarter 2; document the source’s age range, unit and version/revision explicitly. Preserve a missing dimension as unresolved.
- A year-by-year comparison of all nine values, retaining original strings and identifying any comma removal, scaling, rounding or calendar conversion. A wrong province, sex, status, reason or quarter is a nonmatch. A complete match requires all nine year/value pairs under documented equivalent transformations and a supported series version; unexplained differences remain partial.
- Source captures and hashes, acquisition clocks, and the source’s own publication/version assertions separately. Preserve reproducible comparison steps.

An official hostname, an annual average, a different-quarter table or a coincident single value is insufficient. A prior indexed NSO table concerns Quarter 3 of 2003; it supports only the proposed source family and cannot supply missing filters for this series.

## Partial results and limits

A retained source table or reproducible query that establishes only part of the series can be a useful partial result. Missing dimensions, rows or unexplained differences remain open; a partial does not satisfy all criteria.

The native service asserts March 11, 2026 creation times for the three representative objects; the retained API acquisitions occurred on September 5, 2026. These assertions do not independently date the historical full bodies. The years 2013–2021 are data years. This question concerns source and filter recovery, distinct from [chronology issue #54](https://github.com/aadarwal/swarmstatus-bounties/issues/54), which covers historical dating of a separate 13-object set. A matching source would establish content correspondence, without proving copying direction, a particular actor, a unique task or benchmark identity, an official answer fetch, successful use or peer receipt.

Accepted evidence receives public credit; no cash reward is offered. A claim is not exclusive. Follow [the contribution guide](https://github.com/aadarwal/swarmstatus-bounties/blob/main/CONTRIBUTING.md). To announce work, post a comment containing only `/attempt`; describe your scope or findings in a separate comment.


<!-- swarmstatus:roiet-2021-q2-primary-source -->
### Research note: one matching source row (2026-09-05)

The current [NSO Roi Et provincial Labour Force Survey report](https://roiet.nso.go.th/images/64.pdf#page=29) contains **37,842** for **2021, Quarter 2, male, not in the labour force because of studies**. Table 1 on physical PDF page 29 places this value in the estimated population-count section for people aged 15 and over. It matches the pasted `2021: 37,842` line and its stated province, sex, labour-force status, reason and quarter. Same-table controls distinguish Q2 all sexes (74,522), Q2 female (36,680), Q1 male (36,447), the four-quarter male average (36,660), and the Q2 male percentage (9.2).

The retained PDF is 37 pages, 3,581,294 bytes, SHA256 `5e49092ff4cf7c79ee523e919c691a478de2bf17be54666211fd6be758a8130d`, captured on 2026-09-05 at 23:04:39.961564–23:04:44.230401 UTC. Its report data year is B.E. 2564 (2021); its imprint states B.E. 2565 (2022). These source assertions do not independently establish historical PDF or paste availability.

This is a **one-year partial comparison**. The other eight years, their age/unit definitions, and the original series version remain unresolved. The search excerpt at the same URL described a different 2020 table; that excerpt has not been verified against a retained primary document. The separate 2019 candidate returned HTTP 403 and yielded no PDF. Neither indexed candidate fills the remaining gap, and this match does not identify the source actually used to produce the paste. The question remains open with its full acceptance criteria.