[
https://issues.apache.org/jira/browse/AVRO-2211?focusedWorklogId=752065&page=com.atlassian.jira.plugin.system.issuetabpanels:worklog-tabpanel#worklog-752065
]
ASF GitHub Bot logged work on AVRO-2211:
----------------------------------------
Author: ASF GitHub Bot
Created on: 03/Apr/22 21:20
Start Date: 03/Apr/22 21:20
Worklog Time Spent: 10m
Work Description: yanivru commented on code in PR #1597:
URL: https://github.com/apache/avro/pull/1597#discussion_r841278792
##########
lang/csharp/src/apache/main/Schema/EnumSchema.cs:
##########
@@ -103,15 +133,46 @@ internal static EnumSchema NewInstance(JToken jtok,
PropertyMap props, SchemaNam
string doc, string defaultSymbol)
: base(Type.Enumeration, name, aliases, props,
names, doc)
{
- if (null == name.Name) throw new SchemaParseException("name cannot
be null for enum schema.");
+ if (null == name.Name) throw new AvroException("name cannot be
null for enum schema.");
this.Symbols = symbols;
this.symbolMap = symbolMap;
if (null != defaultSymbol && !symbolMap.ContainsKey(defaultSymbol))
- throw new SchemaParseException($"Default symbol:
{defaultSymbol} not found in symbols");
+ throw new AvroException($"Default symbol: {defaultSymbol} not
found in symbols");
Default = defaultSymbol;
}
+ /// <summary>
+ /// Creates symbols map from specified list of symbols.
+ /// Symbol map contains the names of the symbols and their index.
+ /// </summary>
+ /// <param name="symbols">List of symbols</param>
+ /// <returns>Symbol map</returns>
+ /// <exception cref="AvroException">Is thrown if the symbols list
contains invalid symbol name or duplicate symbols</exception>
+ private static IDictionary<string, int>
CreateSymbolsMap(IEnumerable<string> symbols)
+ {
+ IDictionary<string, int> symbolMap = new Dictionary<string, int>();
+ int i = 0;
+ foreach (var symbol in symbols)
+ {
+ if (ValidateSymbolName(symbol))
+ {
+ throw new AvroException($"Invalid symbol name: {symbol}");
+ }
+
+ if (symbolMap.ContainsKey(symbol))
+ {
+ throw new AvroException($"Duplicate symbol: {symbol}");
+ }
+
+ symbolMap[symbol] = i++;
+ }
+
+ return symbolMap;
+ }
+
+ private static bool ValidateSymbolName(string symbol) =>
string.IsNullOrEmpty(symbol) || !Regex.IsMatch(symbol,
"^([A-Za-z_][A-Za-z0-9_]*)$");
Review Comment:
Validate that returns bool. Don't know what I was thinking. Fixed.
Yeah, Regex is not really good at performance (I don't like the readability
either), but since it's a new feature (the parsing code doesn't use this
validation) so we are not degrading performance, and since my guess that
usually schemas are not created in a tight loop (usually created once and
reused), it should be ok. It can be fixed if it does prove to be a problem.
Issue Time Tracking
-------------------
Worklog Id: (was: 752065)
Time Spent: 4h 50m (was: 4h 40m)
> SchemaBuilder equivalent or other means of schema creation
> ----------------------------------------------------------
>
> Key: AVRO-2211
> URL: https://issues.apache.org/jira/browse/AVRO-2211
> Project: Apache Avro
> Issue Type: Wish
> Components: csharp
> Reporter: Dan Stelljes
> Priority: Major
> Labels: pull-request-available
> Time Spent: 4h 50m
> Remaining Estimate: 0h
>
> Currently, the only way to create a schema is via the Schema.Parse method.
> We'd like to be able to build a schema programmatically—would it be possible
> to (1) introduce an analogue to the SchemaBuilder from the Java library or
> (2) expose constructors on Schema and its descendants?
> Would be more than happy to contribute work on this given some direction.
--
This message was sent by Atlassian Jira
(v8.20.1#820001)
