Context. Well-determined scaling relations between X-ray observables and cluster mass are essential for using large cluster samples to constrain fundamental cosmological parameters. Scaling relations between cluster masses and observables, such as the luminositytemperature, mass-temperature, luminosity-mass relations, have been investigated extensively, however the question of whether these relations hold true also for poor clusters and groups remains unsettled. Some evidence supports a "break" at the low end of the group/cluster mass range, possibly caused by the stronger influence of non-gravitational physics on low-mass systems. Aims. The main goal of this work is to test local scaling relations for the low-mass range in order to check whether or not there is a systematic difference between clusters and groups, and to thereby extend this method of reliable and convenient cluster mass determination for future large samples down to the group regime. Methods. We compiled a statistically complete sample of 112 X-ray galaxy groups, 26 of which have usable Chandra data. Temperature, metallicity, and surface brightness profiles were created for these 26 groups, and used to determine the main physical quantities and scaling relations. We then compared the group properties to those of the HIFLUGCS clusters, as well as several other group and cluster samples. Results. We present radial profiles for the individual objects and scaling relations of the whole sample (. Temperature and metallicity profiles behave universally, except for the core regions.and L x − Y x relations of the group sample are generally in good agreement with clusters. The L x − T relation steepens for T < 3 keV, which could point to a larger impact of heating mechanisms on cooler systems. We found a significant drop in the gas mass fraction below < ∼ 1 keV, as well as a correlation with radius, which indicates the ICM is less dominant in groups compared to clusters and the galaxies have a stronger influence on the global properties of the system. In all relations the intrinsic scatter for groups is larger than for clusters, which appears not to be correlated with merger activity but could be due to scatter caused by baryonic physics in the group cores. We also demonstrate the importance of selection effects. Conclusions. We have found some evidence for a similarity break between groups and clusters. However this does not have a strong effect on the scaling relations.