Carlos Alberto Rocha Cardoso created ATLAS-3953: ---------------------------------------------------
Summary: JSON Files from Export API with "?" char for text with special chars Key: ATLAS-3953 URL: https://issues.apache.org/jira/browse/ATLAS-3953 Project: Atlas Issue Type: Bug Components: atlas-core Affects Versions: 2.1.0 Environment: Apache Atlas 2.1.0 embedded HBASE and SOLR Reporter: Carlos Alberto Rocha Cardoso The Export API return a ZIP files with some JSON files describing Atlas Entities and TypeDefs. I am having a issue where some special chars in JSON are being replaced by "?" chars. An Entity name like "Distribuição" was exported in JSON file like "Distribui??o". The special chars "çã" was replaced for the "??" chars. I tried change the exported JSON file encoding and the request header for Export API, but with no success. After analyze the Atlas source code, specialy the splitAndWriteBytes method of the [ZipSink|https://github.com/apache/atlas/blob/cc601d7371fae1dbc16b55d1ca84f06b745700dc/repository/src/main/java/org/apache/atlas/repository/impexp/ZipSink.java] class, I tought if maybe the problem is because the s.getBytes() is returning the JSON string to be written to ZIP with other encode than UTF-8, and maybe setting the the encode like s.getBytes(StandardCharsets.UTF_8) could be a solution. Its my first contact with the Atlas source code, and I'm not a JAVA programmer, so it's only a guess. I saw that it's possible set de default encode to the plataform, or JVM, but how they said in this discussion, perhaps this doesn't work properly in all situations. https://stackoverflow.com/questions/361975/setting-the-default-java-character-encoding -- This message was sent by Atlassian Jira (v8.3.4#803005)