There is a strange behaviour when importing data from MariaDB into R when using MariaDB ODBC 3.1.3 Driver. The table in DB I want to import is:
+----------------------------+---------+-----+------+
| analyse | einheit | MIN | max |
+----------------------------+---------+-----+------+
| ALAT/GPT | U/L | \N | 35 |
| ALAT/GPT | U/L | 1 | 5000 |
| Albumin | g/L | 35 | 55 |
| Albumin | g/L | 1 | 200 |
| Albumin im Urin | mg/L | 0 | 20 |
| Albumin im Urin | mg/L | 0 | \N |
| Albumin/Kreatinin-Quotient | mg/g | 0 | 30 |
| Albumin/Kreatinin-Quotient | mg/g | 0 | 7000 |
| Alkalische Phosphatase | U/L | 36 | 92 |
| Alkalische Phosphatase | U/L | 1 | 3000 |
+----------------------------+---------+-----+------+
If I this into R using MariaDB ODBC 3.1.3 Driver I get this in R:
+------+------------------------------+----------+------+-------+
| row | analyse | einheit | min | max |
+------+------------------------------+----------+------+-------+
| 1 | ALAT/GPT | U/L | NA | 35 |
| 2 | ALAT/GPT | U/L | 0 | 0 | <-
| 3 | Albumin | g/L | 1 | 5000 |
| 4 | Albumin | g/L | 0 | 0 | <-
| 5 | Albumin im Urin | mg/L | 35 | 55 |
| 6 | Albumin im Urin | mg/L | 0 | NA |
| 7 | Albumin/Kreatinin-Quotient | mg/g | 1 | 200 |
| 8 | Albumin/Kreatinin-Quotient | mg/g | 0 | 0 | <-
| 9 | Alkalische Phosphatase | U/L | 0 | 20 |
| 10 | Alkalische Phosphatase | U/L | 0 | 0 | <-
+------+------------------------------+----------+------+-------+
Obviously, in columns min and max couple (0,0) were inserted into rows 2,4,8 and 10 shifting the values of these columns downwards. This happens only with this specific table. If I import other tables from DB to R I got the correct tables.
If I use the MySQL ODBC 8.0 unicode driver it works correctly also with this specific table. Is there a reason for this behaviour and is it possible to avoid it when using this driver?
Environment:
- Windows 10
- R version 3.6.1 (2019-07-05))
- RODBC 1.3-16
- MariaDB 10.4.8
- MariaDB ODBC 3.1 Driver
Update 1: New insights using MariaODBC (Version 3.1.6)
The follow script does compare maria-odbc and mysql-odbc.
library("RODBC")
odbcDataSources()
## -- Two connections
## ma: maria-odbc
## my: mysql-odbc
conn_ma <- odbcConnect("tabl_curr_add_mdo")
conn_my <- odbcConnect("tabl_curr_add")
odbcGetInfo(conn_ma)
odbcGetInfo(conn_my)
head(sqlTables(conn_ma),2)
## +--------------+---------------------+
## | TABLE_SCHEM | TABLE_NAME |
## +--------------+---------------------+
## | codes | ci5v09_ch3_00 |
## | codes | ci5v09_ch3_01 |
## +--------------+---------------------+
head(sqlTables(conn_my),2)
## +---------------+----------------+
## | TABLE_SCHEM | TABLE_NAME |
## +---------------+----------------+
## | tabl_curr_add | arzt |
## | tabl_curr_add | arzt_ctrl_1 |
## +---------------+----------------+
length(unique(sqlTables(conn_ma)$TABLE_CAT))
## 35 databases
length(unique(sqlTables(conn_my)$TABLE_CAT))
## 1 database
Conclusion 1: maria-odbc does connect to all database of the MariaDB instance.
lim_ma <- sqlQuery(conn_ma, "SELECT analyse, min, max FROM tabl_curr_add.labor_limits")
head(lim_ma)
## +--------------------+------+-------+
## | nalyse | min | max |
## +--------------------+------+-------+
## | ALAT/GPT | NA | 35 |
## | ALAT/GPT | 0 | 0 |
## | Albumin | 1 | 5000 |
## | Albumin | 0 | 0 |
## | Albumin im Urin | 35 | 55 |
## | Albumin im Urin | 0 | NA |
## +--------------------+------+-------+
lim_my <- sqlQuery(conn_my, "SELECT analyse, min, max FROM tabl_curr_add.labor_limits")
head(lim_my)
## +--------------------+------+-------+
## | analyse | min | max |
## +--------------------+------+-------+
## | ALAT/GPT | NA | 35 |
## | ALAT/GPT | 1 | 5000 |
## | Albumin | 35 | 55 |
## | Albumin | 1 | 200 |
## | Albumin im Urin | 0 | 20 |
## | Albumin im Urin | 0 | NA |
## +--------------------+------+-------+
Conclusion 2: maria-odbc does corrupt data when imported into R.