CODEWITHSUNDEEP
androidsqliteformattingfloating-point
Tvd
Tvd

Reputation: 4601

Android/SQLite : Saving/Retrieving data of type REAL

I have a SQLDB with table of field REAL. In my layout I have set its field size to allow only 8+2 digits. In my object I have used the data-type of the field as "float".

qs[0] = "CREATE TABLE " + RELATION_TABLE_NAME + "(id TEXT PRIMARY KEY, name TEXT NOT NULL, startBal REAL NOT NULL, currentBal REAL NOT NULL);";

<EditText android:text="" android:id="@+id/relCurrBalTxt" style="@style/EditTextStyle"
 android:inputType="numberDecimal" android:maxLength="11" android:hint="12345678.99" />

I entered value "87654321.99" in my editText and clicked save. It populates the object 

// Send the data to object 
rowData.setValue(2, Float.parseFloat(sbalTxt.getText().toString()));
// In object, this is how I save it
this.setCurrentBalance( ((Float)value).floatValue() );  // value is Object type

// Saving data in ContenteValues to save to DB
// LOG Rcvd from Object while Saving CurrBal: 8.765432E7
values.put("currentBal", new Float(rowData.getValue(3).toString()));

On saving, it directly show the table with the updated data. While showing it in table I use DecimalFormat to make sure it is shown in proper manner :

    field_type = r.getFieldType(field);
// Get Data
str = r.getValue(field).toString();

// Format accordingly
if(field_type.equals(DBRow.DOUBLE_TYPE) || field_type.equals(DBRow.FLOAT_TYPE)) {
        double douValue = Double.parseDouble(str);
            //NumberFormat nf = NumberFormat.getNumberInstance(Locale.ITALY);
            //DecimalFormat df = (DecimalFormat) nf;
            DecimalFormat df = new DecimalFormat();
            df.applyPattern("##,###,###.##");
            df.setGroupingUsed(true);
            str = df.format(douValue);
        //  Log.v("DF", "While Showing in Table : double = " + douValue + " String = " + str);
        } 

((TextView) tr.getChildAt(field)).setText(str);

HERE it showed me the value : 8.765432E7

When I again selected that row to see the values in EditText I see : 87654320.00

How is the value changed ? In other instance also I saved the data as "50009876.99", but somehow it doesn't save .99 and makes that as .00 i.e 50009876.00.

Why things are not working correctly ? Where am I going wrong ?

Any help is highly appreciated.

*EDIT : DecimalFromat code used to display in table * 

        // Format accordingly
    if(field_type.equals(DBRow.DOUBLE_TYPE) || field_type.equals(DBRow.FLOAT_TYPE)) {
        /*
                      WAS USING THIS CODE TILL NOW
    DecimalFormatSymbols dfs = new DecimalFormatSymbols();  
    String decep = String.valueOf(dfs.getDecimalSeparator());
    int index = str.indexOf(decSep); 
    if (index > 0) {
       // If digits after decimal are more than 2
       if (str.substring(index+1).length() > 2) {
        str = str.substring(0, index+3);
    }
    }
    DecimalFormat df = new DecimalFormat();
            df.setGroupingUsed(true);
            str = df.format(Double.parseDouble(str));

            if (addRow)
        create_addTextView(tr, str, true); 
    else 
        ((TextView) tr.getChildAt(field)).setText(str);

            */

                 // DECIMALFORMAT CODE
            double douValue = Double.parseDouble(str);
            DecimalFormat df = new DecimalFormat();
            df.applyPattern("##,###,###.##");
            df.setGroupingUsed(true);
            str = df.format(douValue);

            if (addRow)
       create_addTextView(tr, str, true); 
    else 
      ((TextView) tr.getChildAt(field)).setText(str);

}

Upvotes: 3

Views: 11668

Answers (1)

Boris Strandjev
Boris Strandjev

Reputation: 46963

I think your problem is that you store the numbers as Float. Float does not have very good precision - it actually has only 23 binary digits of precision as seen here. This means that you can not store accurately in float 7 digits before the decimal point or even worse - 7 digits before the point and a couple after. This means that you have incorrectly chosen the type of the balance variable - you can not store in Float 8+2. I would recommend you to change the type to Double, which has significantly larger range of precision.

Upvotes: 6

Related Questions