从昨晚开始,到今天中午之前,一直在纠结时间存储问题,昨晚是纠结时间取出来的问题。
其实我的想法很简单,我就想java.util.Date 存储到 Elasticsearch ,然后从 Elasticsearch 中再取出来的时候,它是个Date ,不需要我任何转换。
但是发现好像不行。
我开始在创建 Mapping 的时候,就是为:
//...省略部分代码.startObject("create_date").field("type","date").field("format","yyyy-MM-dd HH:mm:ss").endObject()//...省略部分代码指定了
Type为Date,并且format为yyyy-MM-dd HH:mm:ss,然后new Date();插入后报错:message [MapperParsingException[failed to parse [create_date]]; nested: IllegalArgumentException[Invalid format: "2016-07-04T03:03:12.616Z" is malformed at "T03:03:12.616Z"];]根据错误提示,我先把时间格式化,然后插入:
result.put("create_date", new SimpleDateFormat("yyyy-MM-dd HH:mm:ss").format(create_date));然后插入OK。后来我看了源码,才恍然大悟。新版本(我不知道从什么版本开始,我以前最开始用的是0.9)值是根据value 的类型来判断。我贴一下。
org.elasticsearch.common.xcontent.XContentBuilder中1248行。PRivate void writeValue(Object value) throws IOException {if (value == null) {generator.writeNull();return;}Class<?> type = value.getClass();if (type == String.class) {generator.writeString((String) value);} else if (type == Integer.class) {generator.writeNumber(((Integer) value).intValue());} else if (type == Long.class) {generator.writeNumber(((Long) value).longValue());} else if (type == Float.class) {generator.writeNumber(((Float) value).floatValue());} else if (type == Double.class) {generator.writeNumber(((Double) value).doubleValue());} else if (type == Byte.class) {generator.writeNumber(((Byte)value).byteValue());} else if (type == Short.class) {generator.writeNumber(((Short) value).shortValue());} else if (type == Boolean.class) {generator.writeBoolean(((Boolean) value).booleanValue());} else if (type == GeoPoint.class) {generator.writeStartObject();generator.writeNumberField("lat", ((GeoPoint) value).lat());generator.writeNumberField("lon", ((GeoPoint) value).lon());generator.writeEndObject();} else if (value instanceof Map) {writeMap((Map) value);} else if (value instanceof Path) {//Path implements Iterable<Path> and causes endless recursion and a StackOverFlow if treated as an Iterable heregenerator.writeString(value.toString());} else if (value instanceof Iterable) {generator.writeStartArray();for (Object v : (Iterable<?>) value) {writeValue(v);}generator.writeEndArray();} else if (value instanceof Object[]) {generator.writeStartArray();for (Object v : (Object[]) value) {writeValue(v);}generator.writeEndArray();} else if (type == byte[].class) {generator.writeBinary((byte[]) value);/* 注意这里:如果是Date类型,就是以字符串输出。如果你跟进去看。代码在下个片段。*/} else if (value instanceof Date) {generator.writeString(XContentBuilder.defaultDatePrinter.print(((Date) value).getTime()));} else if (value instanceof Calendar) {generator.writeString(XContentBuilder.defaultDatePrinter.print((((Calendar) value)).getTimeInMillis()));} else if (value instanceof ReadableInstant) {generator.writeString(XContentBuilder.defaultDatePrinter.print((((ReadableInstant) value)).getMillis()));} else if (value instanceof BytesReference) {BytesReference bytes = (BytesReference) value;if (!bytes.hasArray()) {bytes = bytes.toBytesArray();}generator.writeBinary(bytes.array(), bytes.arrayOffset(), bytes.length());} else if (value instanceof BytesRef) {BytesRef bytes = (BytesRef) value;generator.writeBinary(bytes.bytes, bytes.offset, bytes.length);} else if (value instanceof Text) {Text text = (Text) value;if (text.hasBytes() && text.bytes().hasArray()) {generator.writeUTF8String(text.bytes().array(), text.bytes().arrayOffset(), text.bytes().length());} else if (text.hasString()) {generator.writeString(text.string());} else {BytesArray bytesArray = text.bytes().toBytesArray();generator.writeUTF8String(bytesArray.array(), bytesArray.arrayOffset(), bytesArray.length());}} else if (value instanceof ToXContent) {((ToXContent) value).toXContent(this, ToXContent.EMPTY_PARAMS);} else if (value instanceof double[]) {generator.writeStartArray();for (double v : (double[]) value) {generator.writeNumber(v);}generator.writeEndArray();} else if (value instanceof long[]) {generator.writeStartArray();for (long v : (long[]) value) {generator.writeNumber(v);}generator.writeEndArray();} else if (value instanceof int[]) {generator.writeStartArray();for (int v : (int[]) value) {generator.writeNumber(v);}generator.writeEndArray();} else if (value instanceof float[]) {generator.writeStartArray();for (float v : (float[]) value) {generator.writeNumber(v);}generator.writeEndArray();} else if (value instanceof short[]) {generator.writeStartArray();for (short v : (short[]) value) {generator.writeNumber(v);}generator.writeEndArray();} else {// if this is a "value" object, like enum, DistanceUnit, ..., just toString it// yea, it can be misleading when toString a Java class, but really, jackson should be used in that casegenerator.writeString(value.toString());//throw new ElasticsearchIllegalArgumentException("type not supported for generic value conversion: " + type);}}我们看下这部分:
XContentBuilder.defaultDatePrinter.print(((Date) value).getTime())进去后。看到如下:/*** Prints a millisecond instant to a String.* <p>* This method will use the override zone and the override chronology if* they are set. Otherwise it will use the ISO chronology and default zone.** @param instant millis since 1970-01-01T00:00:00Z* @return the printed result*/public String print(long instant) {StringBuilder buf = new StringBuilder(requirePrinter().estimatePrintedLength());try {printTo((Appendable) buf, instant);} catch (IOException ex) {// StringBuilder does not throw IOException}return buf.toString();}看到这里就明白了吧。他最终的输出方式都是以字符串输出,只是默认的格式是:
1970-01-01T00:00:00Z,也就是默认的UTC格式。我的时间转换结果成:2016-07-04T03:03:12.616Z这里并且有时区的概念,东八区,这里输出的时间少了8个小时。这个得注意。总结了下。最终输出都是
String类型。感觉不友好。我本想的是,我不管存入是怎么样,我取出来得是Date对象就可以了。官网时间(Date)格式说明
关于时间类型说明:https://www.elastic.co/guide/en/elasticsearch/reference/current/date.html
关于时间类型格式化:https://www.elastic.co/guide/en/elasticsearch/reference/current/mapping-date-format.html#strict-date-time
JSON doesn’t have a date datatype, so dates in Elasticsearch can either be:
strings containing formatted dates, e.g. "2015-01-01" or "2015/01/01 12:10:30".a long number representing milliseconds-since-the-epoch.an integer representing seconds-since-the-epoch.Internally, dates are converted to UTC (if the time-zone is specified) and stored as a long number representing milliseconds-since-the-epoch.
Date formats can be customised, but if no format is specified then it uses the default:
"strict_date_optional_time||epoch_millis"This means that it will accept dates with optional timestamps, which conform to the formats supported by strict_date_optional_time or milliseconds-since-the-epoch.
解决方法及问题:
1.时间输出格式,如果是默认 UTC 格式,时间不是我们常用的格式,而且时区问题,少了8个小时。
解决方案:
直接用毫秒值,缺点为不直观。
直接设置format为你想要的格式,比如“yyyy-MM-dd HH:mm:ss”然后存储的时候,指定格式,并且 Mapping 也是指定相同的format。2.存储Date,和取出来也是Dete?
解决方案:
存储的时候利用各种JSON对象,比如 json-lib , fastjson , Jackson , gson 等等。存储的时候就可以用JSON Format一下再存储,然后取出来后,在用JSON.toBean(json,POJO.class),就解决了,这里利用的是相同 JSON 包转成 JSON ,然后又toBean回来,是没问题的,然后 Elasticsearch 也支持存储 JSON 。好了上面观点纯属个人观点。可能存在错误和参杂个人色彩。请勿作为直接参考。错误的地方,请在下面留言。
新闻热点
疑难解答