← All questions

No data file for "data/{env}/users.csv" in env "staging" when the common file exists

sdods run -p shop -e staging -l ui -t @data-driven

✖ DATASET_NOT_FOUND: No data file for "data/{env}/users.csv" in env "staging".
  hint: Tried: /work/shop/projects/shop/data/staging/users.csv, /work/shop/projects/shop/data/common/users.csv

data:
  sources:
    users: { type: csv, path: 'data/{env}/users.csv' }

2 answers

• 15votes✓

  Accepted answer

  Four candidates get tried, in this order:

  1. path with {env} substituted
  2. data/<env>/<basename>
  3. fallback
  4. data/common/<basename>

  The basename is taken from the resolved path, so for data/{env}/users.csv it is users.csv in every one of them. Candidate 4 is data/common/users.csv, which does not exist. logins.csv is never a candidate — nothing in the config mentions that name.

  Two candidates in the hint rather than four because it de-duplicates, and with no fallback set the list collapses.

  Say it explicitly:

  yamlCopy

  data:
    sources:
      users: { type: csv, path: 'data/{env}/users.csv', fallback: data/common/logins.csv }

  The hint prints absolute paths in the order tried, so it is worth reading top to bottom before changing anything — it usually tells you which of the four you meant.

  PVanswered by Priya Venkateshmaintainer29,604 · 23 Mar 2021
• 4votes

  Or rename the file to users.csv and delete the fallback line. Candidate 4 then picks it up for every environment and there is one less thing in the yaml. We keep the rule "the file is named after the dataset" for exactly this reason.

  RGanswered by Rafa Garrido12,996 · 30 Mar 2021