// new processing the mime type out of Uri which may return null in some cases
String mimeType = getContentResolver().getType(uri);
// old processing the mime type out of path using the extension part if new way returned null
if (mimeType == null){mimeType URLConnection.guessContentTypeFromName(path);}

而从资产/文件(注意，在MimeTypeMap中缺少一些情况)。

private String getMimeType(String path) {
    if (null == path) return "*/*";

    String extension = path;
    int lastDot = extension.lastIndexOf('.');
    if (lastDot != -1) {
        extension = extension.substring(lastDot + 1);
    }

    // Convert the URI string to lower case to ensure compatibility with MimeTypeMap (see CB-2185).
    extension = extension.toLowerCase(Locale.getDefault());
    if (extension.equals("3ga")) {
        return "audio/3gpp";
    } else if (extension.equals("js")) {
        return "text/javascript";
    } else if (extension.equals("woff")) {
        return "application/x-font-woff";
    } else {
        // TODO
        // anyting missing from the map (http://www.sitepoint.com/web-foundations/mime-types-complete-list/)
        // reference: http://grepcode.com/file/repo1.maven.org/maven2/com.google.okhttp/okhttp/20120626/libcore/net/MimeUtils.java#MimeUtils
    }

    return MimeTypeMap.getSingleton().getMimeTypeFromExtension(extension);
}

而使用ContentResolver

contentResolver.getType(uri)

当http/https请求

    try {
        HttpURLConnection conn = httpClient.open(new URL(uri.toString()));
        conn.setDoInput(false);
        conn.setRequestMethod("HEAD");
        return conn.getHeaderField("Content-Type");
    } catch (IOException e) {
    }

有时Jeb和Jens的答案不工作并返回null。在这种情况下，我使用跟随解决方案。文件头通常包含类型签名。我阅读了它，并与已知的签名列表进行了比较。

/**
 *
 * @param is InputStream on start of file. Otherwise signature can not be defined.
 * @return int id of signature or -1, if unknown signature was found. See SIGNATURE_ID_(type) constants to
 *      identify signature by its id.
 * @throws IOException in cases of read errors.
 */
public static int getSignatureIdFromHeader(InputStream is) throws IOException {
    // read signature from head of source and compare with known signatures
    int signatureId = -1;
    int sigCount = SIGNATURES.length;
    int[] byteArray = new int[MAX_SIGNATURE_LENGTH];
    StringBuilder builder = new StringBuilder();
    for (int i = 0; i < MAX_SIGNATURE_LENGTH; i++) {
        byteArray[i] = is.read();
        builder.append(Integer.toHexString(byteArray[i]));
    }
    if (DEBUG) {
        Log.d(TAG, "head bytes=" + builder.toString());
    }
    for (int i = 0; i < MAX_SIGNATURE_LENGTH; i++) {

        // check each bytes with known signatures
        int bytes = byteArray[i];
        int lastSigId = -1;
        int coincidences = 0;

        for (int j = 0; j < sigCount; j++) {
            int[] sig = SIGNATURES[j];

            if (DEBUG) {
                Log.d(TAG, "compare" + i + ": " + Integer.toHexString(bytes) + " with " + sig[i]);
            }
            if (bytes == sig[i]) {
                lastSigId = j;
                coincidences++;
            }
        }

        // signature is unknown
        if (coincidences == 0) {
            break;
        }
        // if first bytes of signature is known we check signature for full coincidence
        if (coincidences == 1) {
            int[] sig = SIGNATURES[lastSigId];
            int sigLength = sig.length;
            boolean isSigKnown = true;
            for (; i < MAX_SIGNATURE_LENGTH && i < sigLength; i++) {
                bytes = byteArray[i];
                if (bytes != sig[i]) {
                    isSigKnown = false;
                    break;
                }
            }
            if (isSigKnown) {
                signatureId = lastSigId;
            }
            break;
        }
    }
    return signatureId;
}

signatureId是签名数组中签名的索引。例如,

private static final int[] SIGNATURE_PNG = hexStringToIntArray("89504E470D0A1A0A");
private static final int[] SIGNATURE_JPEG = hexStringToIntArray("FFD8FF");
private static final int[] SIGNATURE_GIF = hexStringToIntArray("474946");

public static final int SIGNATURE_ID_JPEG = 0;
public static final int SIGNATURE_ID_PNG = 1;
public static final int SIGNATURE_ID_GIF = 2;
private static final int[][] SIGNATURES = new int[3][];

static {
    SIGNATURES[SIGNATURE_ID_JPEG] = SIGNATURE_JPEG;
    SIGNATURES[SIGNATURE_ID_PNG] = SIGNATURE_PNG;
    SIGNATURES[SIGNATURE_ID_GIF] = SIGNATURE_GIF;
}

现在我有了文件类型，即使文件的URI没有。接下来我得到mime类型的文件类型。如果你不知道要获取哪种mime类型，你可以在这个表中找到proper。

它适用于许多文件类型。但是对于视频，它不起作用，因为你需要知道视频编解码器来获得mime类型。为了获得视频的mime类型，我使用mediametadatareretrieve。

对于Xamarin Android(来自@HoaLe的回答)

public String getMimeType(Uri uri) {
    String mimeType = null;
    if (uri.Scheme.Equals(ContentResolver.SchemeContent))
    {
        ContentResolver cr = Application.Context.ContentResolver;
        mimeType = cr.GetType(uri);
    }
    else
    {
        String fileExtension = MimeTypeMap.GetFileExtensionFromUrl(uri.ToString());
        mimeType = MimeTypeMap.Singleton.GetMimeTypeFromExtension(
        fileExtension.ToLower());
    }
    return mimeType;
}

public static String getFileType(Uri file)
{
    try
    {
        if (file.getScheme().equals(ContentResolver.SCHEME_CONTENT))
            return subStringFromLastMark(SystemMaster.getContentResolver().getType(file), "/");
        else
            return MimeTypeMap.getFileExtensionFromUrl(file.toString()).toLowerCase();
    }
    catch(Exception e)
    {
        return null;
    }
}

public static String getMimeType(Uri file)
{
    try
    {
        return MimeTypeMap.getSingleton().getMimeTypeFromExtension(getFileType(file));
    }
    catch(Exception e)
    {
        return null;
    }
}

public static String subStringFromLastMark(String str,String mark)
{
    int l = str.lastIndexOf(mark);
    int end = str.length();
    if(l == -1)
        return str;

    return str.substring(l + 1, end);
}

I don't realize why MimeTypeMap.getFileExtensionFromUrl() has problems with spaces and some other characters, that returns "", but I just wrote this method to change the file name to an admit-able one. It's just playing with Strings. However, It kind of works. Through the method, the spaces existing in the file name is turned into a desirable character (which, here, is "x") via replaceAll(" ", "x") and other unsuitable characters are turned into a suitable one via URLEncoder. so the usage (according to the codes presented in the question and the selected answer) should be something like getMimeType(reviseUrl(url)).

private String reviseUrl(String url) {

        String revisedUrl = "";
        int fileNameBeginning = url.lastIndexOf("/");
        int fileNameEnding = url.lastIndexOf(".");

        String cutFileNameFromUrl = url.substring(fileNameBeginning + 1, fileNameEnding).replaceAll(" ", "x");

        revisedUrl = url.
                substring(0, fileNameBeginning + 1) +
                java.net.URLEncoder.encode(cutFileNameFromUrl) +
                url.substring(fileNameEnding, url.length());

        return revisedUrl;
    }