初始化结果 warning怎么进行下一步

[jifeisong]

每次初始化都是这个结果，daily_bars有warning，怎么进入下一步

{
"run_id": "cee63a18-c77b-414e-9455-895f6dbfa328",
"status": "failed",
"resumed": true,
"retry": {
"run_id": "cee63a18-c77b-414e-9455-895f6dbfa328",
"status": "failed",
"retried": 3,
"missing_steps": [],
"missing_steps_unresolved": [
"trading_status_derive"
],
"stale_marked_failed": 0,
"batch_timeout": {
"running_to_stale": 0,
"stale_to_failed": 0
},
"results": [
{
"step": "daily_bars",
"status": "warning",
"elapsed": 1.0702601809753105,
"rows_read": 0,
"rows_written": 0,
"metrics": {
"rows_read": 0,
"rows_written": 0,
"source_failures": {
"rate_limited": 0,
"source_empty": 0,
"proxy_failed": 0,
"direct_failed": 0,
"circuit_open": 0,
"transport_error": 0,
"http_error": 0
}
},
"context_updates": {
"audit_findings": [
{
"dataset": "daily_bars",
"severity": "warning",
"check": "daily_bars_source_ownership",
"message": "generic=0, delegated_delisted=102, expected_no_data=0, placeholder=0, probe=0, negative_cached=0, unknown=0, delegated_complete=False",
"start": "2023-09-30",
"end": "2026-09-30"
}
],
"daily_bars_ownership": {
"generic": 0,
"delegated_delisted": 102,
"expected_no_data": 0,
"placeholder": 0,
"probe": 0,
"negative_cached": 0,
"unknown": 0,
"delegated_complete": false
}
},
"delegated_symbols": 102
},
{
"step": "trading_status",
"status": "warning",
"elapsed": 4081.505952592008,
"rows_read": 260979,
"rows_written": 260979,
"completed_symbols": 800,
"expected_symbols": 5560,
"source_results": [
{
"rows_read": 260979,
"rows_written": 260979,
"scope_id": "13073e0549c4c2db12e40042821d59a52d7d5dd7217185958eb41863fb768ede",
"evidence_version": 2,
"checkpoint": "/home/hadoop/X/cnequity/data/cnequity/meta/state/historical_st_evidence/v2/13073e0549c4c2db12e40042821d59a52d7d5dd7217185958eb41863fb768ede.json",
"completed_symbols": 800,
"expected_symbols": 5560,
"source": "baostock",
"status": "warning",
"failed_symbols": 0,
"deferred_symbols": 4760,
"context_updates": {
"audit_findings": [
{
"dataset": "trading_status",
"severity": "warning",
"code": "baostock_st_backfill_incomplete",
"message": "800/5560 symbols have baostock ST evidence: 4760 not yet swept (paused at 400 symbols/run). Resume with cne backfill trading_status \u2014 the checkpoint continues from here, and this run's rows are already staged."
}
]
}
}
],
"unsupported_symbols": 348,
"unsupported_exchanges": [
"BJ"
],
"failed_symbols": 0,
"context_updates": {
"audit_findings": [
{
"dataset": "trading_status",
"severity": "warning",
"code": "baostock_st_backfill_incomplete",
"message": "800/5560 symbols have baostock ST evidence: 4760 not yet swept (paused at 400 symbols/run). Resume with cne backfill trading_status \u2014 the checkpoint continues from here, and this run's rows are already staged."
}
]
}
},
{
"step": "compact",
"status": "warning",
"elapsed": 4.702996815030929,
"rows_read": 31743,
"rows_written": 31743,
"context_updates": {
"compact_skipped_datasets": [
{
"dataset": "trading_status",
"incomplete_batches": 2
},
{
"dataset": "daily_bars",
"incomplete_batches": 1
}
]
}
}
],
"retry_passes": 1,
"retry_exhausted": 0
},
"phases": [
{
"phase": "phase4_partial_compact",
"status": "warning",
"results": [
{
"step": "compact",
"status": "warning",
"elapsed": 4.813542090007104,
"rows_read": 31743,
"rows_written": 31743,
"context_updates": {
"compact_skipped_datasets": [
{
"dataset": "trading_status",
"incomplete_batches": 2
},
{
"dataset": "daily_bars",
"incomplete_batches": 1
}
]
}
}
]
}
]
}

[Arthur031221]

Your run has two unfinished jobs. The trading_status warning is a deliberate pause of the slow Baostock historical ST scan: 800 of 5,560 symbols have evidence and 4,760 have not been swept. failed_symbols: 0 means this warning is not reporting failed symbol requests. Separately, daily_bars delegated 102 delisted symbols and reports delegated_complete=False. The missing trading_status_derive phase and skipped compact are consequences of the incomplete initialization.

I checked the current CLI source and its initialization guide. This short source check confirms the important distinction between the bounded init pass and the explicit backfill:

$ python3 - <<'PY'
from pathlib import Path
loader = Path('src/cnequity/config/loader.py').read_text()
backfill = Path('src/cnequity/cli/backfill_cmds.py').read_text()
print('init defaults to 400 symbols/run:', 'st_history_symbols_per_run: int = 400' in loader)
print('explicit backfill removes limit:', 'cfg.st_history_symbols_per_run = 0' in backfill)
print('not yet swept:', 5560 - 800)
PY
init defaults to 400 symbols/run: True
explicit backfill removes limit: True
not yet swept: 4760

Keep the existing data directory and checkpoint. From the same installation and with the same config, finish the two distinct backfills, then resume initialization:

cne backfill trading_status --config configs/cnequity.toml
cne backfill daily_bars --profile delisted --start 2023-09-30 --config configs/cnequity.toml
cne init --resume --config configs/cnequity.toml
cne status --datasets --config configs/cnequity.toml

The first command removes the 400-symbol limit and resumes the ST checkpoint; allow a long window because it scans symbols one at a time under source pacing. The second handles the separate delisted-bar coverage indicated by delegated_complete=False. --start 2023-09-30 matches the start of the warning's reported interval. The initialization guide documents the cap, checkpoint resume, and init --resume; the CLI reference documents the delisted profile. Recheck the final status before treating the lake as complete, since source coverage can still limit some symbols.

[rootSunc]

在 v0.12.0 上，Arthur031221 给的步骤是对的（先跑 cne backfill trading_status，再补退市日线，然后 resume init）。我在版本v0.13.0后，这些手动步骤都压进一条init命令中。

[rootSunc]

新版本v13.0.0已经发布

v0.13.0 的变化：

历史 ST 扫描不再属于 init，也不再是初始化完成的条件。init 获取当前交易状态快照，日线发布后再派生窗口内的历史停牌。
init 自动恢复窗口内已知退市股票的日线，不再需要另外运行 --profile delisted。
来源受限时，经过校验的部分结果照常发布，命令以 warning / degraded 结束并返回 0，不再报 failed。缺口会记下来供之后补数。
升级后怎么做： 不要删除现有的 data/ 目录，已抓取的数据和 checkpoint 都会保留。

pip install -U cnequity      # 源码安装的话：git pull && pip install -e .
cne config upgrade           # 把新版调度步骤补进你的配置，会先备份原文件
cne init                     # 重跑同一条 init，会自动续跑之前未完成的 run
cne check                    # 验收：新鲜度、覆盖、审计结果，最后给出结论

如果你当初用 --config 指定了配置文件，以上每条命令都加上同样的 --config。之前回复里分步执行 backfill trading_status、backfill daily_bars --profile delisted、init --resume 的做法，现在都不需要了。

只有需要历史 ST 研究时，再单独运行：
cne backfill trading_status --start 2016-01-01 --end YYYY-MM-DD
这条命令没有 400 只的上限，会从已有的 checkpoint 继续。它逐只证券、按来源节奏扫描，需要较长时间；Baostock 不提供北交所的历史 ST。

详见初始化、范围与续跑。升级后如果还有问题，请贴上 cne check 的输出。